DEV Community: Hemapriya Kanagala

🐾 DEV Pet Corner #4: Show Us Your Pets!

Hemapriya Kanagala — Sun, 01 Nov 2026 17:53:50 +0000

TL;DR

DEV Pet Corner is back for another Sunday! It's a weekly series where we take a little break from tech and share our love for animals.

Share pictures, stories, and memories of your pets, animals you've met, or adorable animals you follow online. Share as many pictures and links as you'd like! You don't need to own a pet to join in. If animals make you smile, you're welcome here ❤️

❤️ Thank you for the love
🐶 Show us your animal friends!
🐾 See you next Sunday!

❤️ Thank you for the love!

Hey everyone 👋 It's Sunday again, and DEV Pet Corner is back for its fourth edition!

First of all, thank you so much to everyone who joined the previous edition ❤️ I loved seeing all your pictures, reading your stories, and getting to know your animal friends. It's always so lovely to see everyone come together and share something they love. I'm really happy this little corner of DEV is bringing us together!

Thank you for taking the time to share your pictures and a little piece of your lives with us. It really means a lot to me!

If you missed the previous edition, you can still meet all the cuties here 👉 DEV Pet Corner #3

Now, let's do this again 😄

🐶 Show us your animal friends!

I'd love to see your pets, whether they're the ones waiting for you at home, the ones you grew up with, or the ones who have a special place in your heart.

And you don't need to have a pet of your own to join in! Maybe you've met a friendly dog on a walk, spotted a cat in your neighborhood, or come across an animal on campus that made you stop for a moment. Perhaps you follow an adorable animal account on Instagram or have a favorite animal video you keep going back to.

If it involves an animal you love or something that makes you smile, I'd love to see it 😃

Share pictures, funny moments, sweet memories, little stories, or links to your favorite animal posts. Share as many pictures as you'd like. There's no limit!

If you're sharing someone else's content, please link to the original post and respect the creator's wishes about sharing their work.

A little note for returning readers: You might notice this section looks familiar 😄 I'll keep it in future editions, too, so new readers can easily see what they can share without going back to previous posts. As always, feel free to join in with anything you'd like or simply enjoy all the cuteness in the comments!

🐾 See you next Sunday!

I'm so happy this little series is bringing people together over something we love. Thank you for being part of it and making each edition special ❤️

As always, let's keep the comments friendly and welcoming to everyone. I'm looking forward to seeing your animal friends and enjoying all the cuteness with you!

Thank you for being here. See you next Sunday for DEV Pet Corner #5!

Dev Opportunity Radar #23: $500K Mozilla AI Incubator, $100K+ Kiroween Hackathon, and Wispr Flow Hiring Challenge

Hemapriya Kanagala — Fri, 30 Oct 2026 18:28:18 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features Mozilla's Human-Scale AI Incubator, offering up to $500,000 in potential funding for selected projects; Kiroween 2026, a Kiro-powered hackathon with $100,000+ in prizes; and the Wispr Flow Hiring Challenge, where candidates can build a creative portfolio using voice to explore hiring opportunities with its India growth team. I'm also sharing the Hugging Face AI Agents Course, a free resource with hands-on lessons, assignments, and certification options for people interested in learning to build AI agents.

If you're new to the series, you can browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

This week's 🌟 Community Finds section features an opportunity shared by La Shara Cordero (@earlgreyhot1701d). Thank you for sharing the AdaL Ambassador Program with the community. I always enjoy seeing readers help others discover opportunities they might otherwise have missed, and I hope this section continues to grow.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, attended an event, joined a community, completed a program, built something, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, event, or anything else you think deserves more attention, feel free to share it in the comments.

If I feature one of your 🌟 Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Mozilla Human-Scale AI Incubator
- 📌 Kiroween 2026
- 📌 Wispr Flow Hiring Challenge
📚 Resources Worth Checking Out
- Hugging Face AI Agents Course
🌟 Community Finds
- The AdaL Ambassador Program
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline / Status
Mozilla Human-Scale AI	Mozilla Foundation	AI Incubator / Funding	Global applicants (subject to eligibility)	November 20, 2026
Kiroween 2026	Kiro	AI Hackathon	Online	December 4, 2026
Wispr Flow Hiring Challenge	Wispr Flow	Hiring Challenge	India / Bangalore or willing to travel	November 4, 2026

Resource Highlight

Resource	Why Check It Out
Hugging Face AI Agents Course	A free, hands-on course covering AI agent fundamentals, popular frameworks, practical assignments, challenges, and certification options.

Community Finds

Shared By	Find	Type	Location / Format
La Shara Cordero (@earlgreyhot1701d)	AdaL Ambassador Program	Ambassador Program	Worldwide / Cities & Campuses

📝 A quick note: I spend a lot of time researching and reviewing the opportunities I feature in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications or submissions.

As the Radar has grown, there are now quite a few opportunities spread across the different editions. I used to keep all the still-open opportunities here, but I don't think repeating a long list every week makes the editions easier to read.

So I keep this section to the two most recent editions as a quick reminder of opportunities you may have missed. Older opportunities are still available on the Dev Opportunity Radar website, where you can browse the collection and use the filters to find something specific.

I've already covered these opportunities in detail, so I won't repeat everything here. If any of them catch your attention, you can go back to the original edition for the full overview, eligibility details, and application links.

Opportunity	Organization	Type	Format	Deadline / Status	Featured In
Basis Fellowship	Basis	Fellowship	Not specified	Rolling applications	Edition #22
More Than a Hack 2027	Mobile World Capital Barcelona, Barça Innovation Hub & Metrica Sports	Football Analytics Hackathon	Online challenges + final in Barcelona	November 10, 2026 (registration)	Edition #22
Perplexity Research Fellowship	Perplexity	AI Research Fellowship	In person: San Francisco, Palo Alto, or New York City	Rolling applications; priority deadline passed	Edition #22
MATS Residency: AI Safety Research	MATS	Research Residency	Hybrid: London, Berkeley, or Washington, DC	October 31, 2026	Edition #21
Y Combinator Winter 2027 Batch	Y Combinator	Startup Accelerator & Funding	In person: San Francisco, USA	November 2, 2026	Edition #21

Please check the original editions for complete details and confirm that each opportunity is still accepting applications or submissions before proceeding.

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Mozilla Human-Scale AI Incubator

Who it's for: Teams and builders who already have working AI technology deployed with real users, particularly those developing smaller, resource-efficient, privacy-conscious, and purpose-built AI systems.

What it's about:

The Mozilla Foundation is looking for projects that take a different approach to building AI. Instead of focusing on larger models, more compute, and systems designed to do everything, Human-Scale AI focuses on building technology around the specific people and problems it is meant to serve.

The call is built around three principles: Small, Intentional, and Tailored. In simple terms, Mozilla wants to support AI systems that use resources efficiently, give people more control over their data and technology, and are designed for specific communities and real-world needs.

This is not an early-stage idea or prototype grant. Projects must already have working technology and some evidence that people value or use what they're building. Mozilla is particularly interested in projects serving underserved communities and those offering alternatives to the resource-intensive and centralized AI systems we see today.

Selected projects will receive $300,000 in funding and 12 months of support, including technical coaching, mentorship, and access to Mozilla's wider network. Projects that demonstrate strong progress may be eligible for up to another $200,000 in funding and two additional years of support, bringing the potential total to $500,000 over three years. The funding is non-dilutive, so applicants don't give up equity.

One thing worth checking carefully before applying is eligibility. Mozilla is looking for primarily technical projects where AI is a core part of the solution. Generic chatbot applications, standalone research or policy projects, and ideas without working technology are not what this call is looking for. Projects must also demonstrate a charitable purpose and meet Mozilla's funding eligibility requirements.

If you're already building something in this space and have people using it, this could be worth exploring.

Funding: $300,000 initially, with up to $200,000 in additional funding for eligible projects demonstrating strong progress

Total potential support: Up to $500,000 over three years

Program duration: 12 months initially, with potential for two additional years of support

Cohort start: April 2027

Number of projects: 7

Application deadline: November 20, 2026, at 11:59 p.m. Pacific Standard Time (UTC-8)

Format: Incubator program with funding, coaching, and cohort support

Equity: No equity requirement

About Human-Scale AI | Call for proposals and application | Eligibility guidelines

📌 Kiroween 2026

Who it's for: Developers, students, and builders interested in experimenting with AI-powered development, building software applications, and exploring tools such as AI agents, MCP, and spec-driven development.

What it's about:

Kiroween 2026 is Kiro's annual Halloween-themed hackathon, where participants are invited to build a new software application using Kiro. This year's event has a prize pool of $100,000+, with several categories that give participants different ways to approach the challenge.

There are four main categories to choose from. Resurrection is about bringing an old or obsolete technology back to life using modern approaches. Frankenstein challenges you to combine seemingly unrelated technologies into something useful. Skeleton Crew focuses on building a flexible code template and demonstrating it through two different applications. And if you enjoy working on the visual side of development, Costume Contest is about creating an application with a polished and memorable interface.

You can participate individually or in a team of up to three members. The important thing to note is that projects must be new and built during the hackathon's submission period. Projects started before October 30, 2026, including entries from Kiroween 2025, aren't eligible.

There are also some specific submission requirements. Your project needs to be a working application built using Kiro, and you'll need to submit a public GitHub repository with an open-source license and the .kiro directory included. You'll also need a working application link, a project description, a demo video of no more than three minutes, and an explanation of how you used Kiro during development.

Beyond the main categories, there are additional prizes for student projects, startup projects, Kiro Crew applications, and other contributions such as blog posts and social media posts.

One eligibility detail worth checking before registering: participants must be above the legal age of majority in their country of residence, and certain countries and territories are excluded under the official rules. I'd recommend checking the full rules to confirm that your location is eligible before getting started.

Prize pool: $100,000+ in cash and additional prizes

First place: $30,000 | Second place: $20,000 | Third place: $10,000

Categories: Resurrection, Frankenstein, Skeleton Crew, and Costume Contest

Team size: Individual participation or teams of up to 3

Format: Online

Project requirement: New working software application built using Kiro

Submission period: October 30–December 4, 2026

Deadline: December 4, 2026, at 2:00 p.m. Pacific Time (December 5 at 3:30 a.m. IST)

If you're planning to participate, take some time to read the submission requirements and choose a category before you start building. There are quite a few details to get right, especially around the GitHub repository, .kiro directory, and demo video.

Join Kiroween 2026

📌 Wispr Flow Hiring Challenge

Who it's for: Students, aspiring growth professionals, and experienced professionals interested in joining Wispr Flow's India growth team. Internships and full-time opportunities are available, depending on experience.

What it's about:

Wispr Flow is running a hiring challenge to find creative people for its growth team in India. Instead of asking candidates to submit just a resume and application, they're asking them to build a new portfolio using Wispr Flow and a tool such as Emergent.

The idea is to show how you think, what you can create, and how you use voice as part of your building process. You can build an interactive portfolio, a game, a quiz, a mock interview app, an interactive timeline, or something entirely different. The focus is on creativity, how you communicate your ideas, and what you can take ownership of.

There are opportunities for both interns and experienced full-time candidates. For internships, having at least one year of experience is preferred, but it's not a strict requirement, and students are welcome to apply. Full-time roles are aimed at people who have previously owned or grown channels at tech startups, particularly in areas such as influencer marketing and partnerships.

A few things to keep in mind before applying: this is specifically for the India growth team, and candidates should be based in Bangalore or willing to travel. The team ideally wants people who can start in November, although they're open to waiting for the right candidate. Students graduating later are also welcome.

You'll need to submit a new portfolio built for this challenge, a publicly accessible link, and a video explaining what you created, why you chose it, and how voice helped you build it. Existing portfolios created before the challenge aren't accepted. Wispr Flow is providing Emergent credits to help participants get started, although you're free to use another vibe-coding tool.

Opportunity: Internship and full-time hiring opportunities

Team: Growth, India

Location: Bangalore or willing to travel

Preferred start: November 2026, with flexibility for the right candidate

Challenge dates: October 28–November 4, 2026

Shortlist notification: November 5, 2026

Hiring announcement: November 20, 2026

Compensation: Top-of-market pay (exact figures not specified)

Additional detail: The submission form also asks candidates to select a $3,000 work setup pick.

This one is a little different from the usual hackathon or fellowship. If you're interested in growth roles and enjoy building creative things, it could be worth exploring, even if you don't usually think of yourself as someone who builds portfolios.

Explore the Wispr Flow Hiring Challenge

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

Hugging Face AI Agents Course

Who it's for: Developers and learners with basic Python and LLM knowledge who want to understand AI agents and learn how to build them.

What it's about:

Hugging Face has a free AI Agents Course that takes you through the fundamentals of AI agents and helps you build your own using popular libraries and tools.

The course starts with the basics, covering how agents work, including tools, thoughts, actions, and observations. From there, it moves into frameworks such as smolagents, LangGraph, and LlamaIndex, with hands-on sections where you can experiment with building agents in preconfigured Hugging Face Spaces.

You'll also work on use case assignments, share your agents on the Hugging Face Hub, and participate in challenges where you can compare your agent's performance with other learners on a leaderboard. There are additional units covering topics such as fine-tuning LLMs for function calling, agent observability and evaluation, and even building agents that play Pokémon games.

One thing I found useful is that you can choose how you want to follow the course. You can audit it and learn at your own pace, or complete the required assignments and challenges to earn one of the certificates. The certification process is completely free, and there is no deadline to complete it.

You should have basic knowledge of Python and LLMs before getting started. The recommended pace is around 3–4 hours per week, with each chapter designed to take approximately one week.

Format: Free online course with theory, hands-on exercises, and challenges

Prerequisites: Basic Python and LLM knowledge

Libraries covered: smolagents, LangGraph, and LlamaIndex

Additional topics: Function calling, agent observability, evaluation, and game-playing agents

Certification: Free, with fundamentals and course completion certificates

Deadline: No deadline for certification

If you've been wanting to understand how AI agents actually work and try building some yourself, this is a good resource to explore.

Explore the Hugging Face AI Agents Course | GitHub Repository

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

Sometimes, a reader comes across something, tries it out themselves, and then brings it here for others to explore. That's exactly what happened with this week's find.

The AdaL Ambassador Program

Shared by La Shara Cordero (@earlgreyhot1701d).

@earlgreyhot1701d shared an opportunity from AdaL, where they're looking for ambassadors from cities and campuses around the world to help organize workshops and hackathons.

The idea is to have people from different communities help bring AdaL workshops and hackathons to their own cities or campuses. You don't need to be from a particular location. They're looking for people from anywhere in the world.

As an ambassador, you'll receive mentorship on organizing events, building your personal brand, improving your employability skills, and learning about product building and go-to-market (GTM). AdaL also mentions subscriptions for ambassadors and event participants, merchandise such as T-shirts and laptop stickers, reimbursement for food and venue expenses, and opportunities to have your events featured on its social media channels and email list.

There are a few things to keep in mind before reaching out. They're looking for people who can demonstrate AdaL's products, including its CLI and desktop features, and who have experience with communities, such as university societies, developer groups, or AI communities. You'll also need to be available to organize or co-host events regularly, with the first event expected by the end of November 2026.

I wanted to include this because community involvement can take different forms. Sometimes it's about attending events, and sometimes it's about taking the initiative to organize something for others. If you're already part of a campus or developer community and enjoy bringing people together, this might be something to look into.

Organization: AdaL

Role: Community Ambassador

Location: Worldwide, including cities and campuses

What you'll do: Organize or co-host workshops and hackathons

Expected commitment: Monthly events

First event: Expected by the end of November 2026

Contact: Jax Koh (jax_koh) on Discord

🔗 Explore AdaL

Thank you to La Shara Cordero (@earlgreyhot1701d) for thinking of the Radar and sharing this opportunity with us. I really appreciate readers taking the time to pass along things they've discovered themselves 💙

I'd love for this section to keep growing.

If you've come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, feel free to share it in the comments.

If I feature it in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

If you discovered an opportunity through the radar, applied to something, joined a community, attended an event, or simply found a resource you hadn't seen before, I'd genuinely love to hear about it.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

If you'd like to share an update, feel free to leave a comment on this edition. With your permission, I may feature it in a future 💙 Reader Updates section and tag you so other readers can celebrate your journey too.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

Every week, I spend time searching for opportunities, resources, and communities that I think deserve a little more attention. But one of my favorite parts of this series isn't the research. It's seeing what happens after an edition is published.

Seeing someone discover an opportunity, apply to a program, share a resource, suggest a Community Find, or come back to tell us what they learned reminds me why I started Dev Opportunity Radar in the first place.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Whether you've been reading since the very first edition or this is your first time here, thank you for being part of the journey. Every comment, suggestion, and shared opportunity helps make this series better than I could build on my own.

If you ever come across an opportunity, resource, event, community, or anything else you think deserves more attention, I'd love for you to share it in the comments. And if Dev Opportunity Radar helped you discover something exciting, I'd love to hear that too.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

If this is your first time discovering the series, you can explore every edition, browse opportunities by category, discover Community Finds, and catch up on Reader Updates on the Dev Opportunity Radar website.

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

🐾 DEV Pet Corner #3: Show Us Your Pets!

Hemapriya Kanagala — Sun, 25 Oct 2026 18:28:32 +0000

TL;DR

DEV Pet Corner is back for another Sunday! It's a weekly series where we take a little break from tech and share our love for animals.

Share pictures, stories, and memories of your pets, animals you've met, or adorable animals you follow online. Share as many pictures and links as you'd like! You don't need to own a pet to join in. If animals make you smile, you're welcome here ❤️

❤️ Thank you for the love
🐶 Show us your animal friends!
🐾 See you next Sunday!

❤️ Thank you for the love!

Hey everyone 👋 It's Sunday again, and DEV Pet Corner is back for its third edition!

First of all, thank you so much to everyone who joined the previous edition ❤️ I loved seeing your pictures, reading your stories, and getting to know more of your animal friends. It's so nice seeing familiar faces return with more pictures, and I love that we're making this little corner of DEV a place where we can share something we love.

Thank you for taking the time to share your pictures and a little piece of your lives with us. It really means a lot to me!

If you missed the previous edition, you can still meet all the cuties here 👉 DEV Pet Corner #2

Now, let's do this again 😄

🐶 Show us your animal friends!

I'd love to see your pets, whether they're the ones waiting for you at home, the ones you grew up with, or the ones who have a special place in your heart.

If it involves an animal you love or something that makes you smile, I'd love to see it 😃

Share pictures, funny moments, sweet memories, little stories, or links to your favorite animal posts. Share as many pictures as you'd like. There's no limit!

If you're sharing someone else's content, please link to the original post and respect the creator's wishes about sharing their work.

🐾 See you next Sunday!

I'm so happy this little series is bringing people together over something we love. Thank you for being part of it and making each edition special ❤️

As always, let's keep the comments friendly and welcoming to everyone. I'm looking forward to seeing your animal friends and enjoying all the cuteness with you!

Thank you for being here. See you next Sunday for DEV Pet Corner #4!

Dev Opportunity Radar #22: $50K Basis Fellowship, Perplexity AI Research Fellowship, and FC Barcelona Analytics Hackathon

Hemapriya Kanagala — Fri, 23 Oct 2026 18:29:07 +0000

TL;DR

Welcome back to Dev Opportunity Radar!

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the Basis Fellowship, offering $50,000 to selected fellows; More Than a Hack 2027, a football analytics hackathon using real FC Barcelona match data; and the Perplexity Research Fellowship for early-career AI researchers. I'm also sharing Harness Engineering 101, a free online book that explores how AI agents work, from tool calling and context management to coding agents and retrieval-augmented generation.

If you're new to the series, you can browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, joined a community, completed a program, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section, with your permission.

And if you've come across an opportunity, resource, community, program, or event that deserves more attention, feel free to share it in the comments.

If I feature one of your Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Basis Fellowship
- 📌 More Than a Hack 2027
- 📌 Perplexity Research Fellowship
📚 Resources Worth Checking Out
- Harness Engineering 101
🌟 Community Finds
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline / Status
Basis Fellowship	Basis	Fellowship	Location not specified	Rolling applications
More Than a Hack 2027	Mobile World Capital Barcelona, Barça Innovation Hub & Metrica Sports	Football Analytics Hackathon	Online challenges + final in Barcelona	November 10, 2026 (registration)
Perplexity Research Fellowship	Perplexity	AI Research Fellowship	In person: San Francisco, Palo Alto, or New York City	Rolling applications; priority deadline passed

Resource Highlight

Resource	Why Check It Out
Harness Engineering 101	A practical, chapter-by-chapter introduction to building AI agents, with a small Python implementation that develops as new concepts are introduced.

📝 A quick note: I spend a lot of time researching and reviewing the opportunities I feature in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications or submissions.

I've already covered the opportunities below in detail, so I won't repeat everything here. If any of them catch your attention, you can go back to the original edition for the full overview, eligibility details, and application links.

Opportunity	Organization	Type	Format	Deadline	Featured In
MATS Residency: AI Safety Research	MATS	Research Residency	Hybrid: London, Berkeley, or Washington, DC	October 31, 2026	Edition #21
Y Combinator Winter 2027 Batch	Y Combinator	Startup Accelerator & Funding	In person: San Francisco, USA	November 2, 2026	Edition #21
PayPal AI Hackathon	PayPal	AI Hackathon	Online, worldwide	November 13, 2026	Edition #20
ARC Prize 2026	ARC Prize	AI Competition & Research	Online / Kaggle	November 2, 2026 (papers: November 8)	Edition #20

Please check the original editions for complete details and confirm that each opportunity is still accepting applications or submissions before proceeding.

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Basis Fellowship: $50,000 for Selected Fellows

Who it's for: People working on interesting or unconventional projects, including engineers, scientists, founders, operators, and others pursuing work across different fields.

What it's about:

The Basis Fellowship offers $50,000 to selected fellows, along with access to a network of people working across different areas.

Rather than focusing on one particular industry or type of project, the fellowship is interested in people doing a wide range of work. The examples on its website include an electrical setup, a thermoelectric cloud chamber, and a neurostimulation machine. The idea is to bring together people who might be working on very different things but could benefit from finding one another.

The fellowship doesn't require equity, reporting, or restrictions on how the money is spent. Applicants can submit a portfolio and other useful information, and the application asks them to answer the questions honestly.

People from different backgrounds can apply, whether they're engineers, scientists, operators, or working in another field. Applications are accepted on a rolling basis, and an application doesn't expire if someone isn't selected immediately. Applicants can send updates when something new happens.

There's also a nomination option. If someone you nominate is selected, you receive $1,000. The website also mentions a discretionary $3,000 award for a strong submission that isn't selected.

Funding: $50,000 for selected fellows
Equity: None
Eligibility: Open to people across different fields and backgrounds
Application process: Submit a portfolio and other relevant information
Applications: Rolling basis
Deadline: No fixed deadline stated
Nomination reward: $1,000 if your nominee is selected
Additional award: A discretionary $3,000 for a strong submission that isn't selected

Fellowship overview and application | Nominate someone

📌 More Than a Hack 2027: Football Analytics Hackathon

Who it's for: Football analysts, data analysts and scientists, software developers, and tech enthusiasts interested in using data and technology to understand football.

What it's about:

More Than a Hack 2027 is a football analytics hackathon organized by Mobile World Capital Barcelona, Barça Innovation Hub, and Metrica Sports. Participants work with real FC Barcelona match data to develop analytical solutions related to the game.

The hackathon has two online challenge stages, followed by an in-person final pitch at Talent Arena 2027 in Barcelona.

The Initial Challenge asks teams to estimate team possession for every frame of a football match using provided automatic tracking data and homography. Participants don't need to build player detection, identification, appearance, or calibration systems from scratch. Submissions are automatically evaluated and ranked.

The top 20 teams advance to the Innovation Challenge. In this stage, teams can explore analytical or tactical questions using a dataset of 10 professional matches. They can experiment with different approaches to analysis, modeling, interpretation, and generating insights, with specialized technical mentoring available.

The top five teams from this stage will be invited to present their projects at Talent Arena 2027 in Barcelona. Travel and accommodation expenses for the final pitch are covered. The listed prizes also include official FC Barcelona shirts, Barça Innovation Hub merchandise, and an exclusive FC Barcelona experience.

Teams must have 3 to 5 members. Teams with fewer than three members won't be eligible for the Innovation Challenge or the Final Pitch.

Registration is open until November 10, 2026. The Initial Challenge is already underway, so anyone interested should review the schedule and registration requirements before joining.

Format: Two online challenges followed by an in-person final pitch
Team size: 3–5 members
Data provided: FC Barcelona match videos, automatic tracking data, and automatic event data
Registration: October 1–November 10, 2026
Initial Challenge: October 20–November 17, 2026
Innovation Challenge: November 26, 2026–January 14, 2027
Final Pitch: March 2, 2027, at Talent Arena in Barcelona
Finalists' travel: Travel and accommodation expenses covered for the final pitch

More Than a Hack 2027 registration and details

📌 Perplexity Research Fellowship: AI Research Program

Who it's for: Early-career researchers, engineers, and analysts from technical or quantitative disciplines, including fields such as physics, mathematics, cognitive science, and quantitative finance.

What it's about:

The Perplexity Research Fellowship is a full-time research program for early-career talent interested in advancing AI research. Fellows join Perplexity's R&D teams and work on research areas such as multi-agent planning, continual learning, pre-training, and human-AI interaction.

The program has two tracks:

Predoctoral Research Fellow: For exceptional early-career talent from technical or quantitative disciplines.
Postdoctoral Research Fellow: For researchers who have recently completed a PhD in a technical or quantitative discipline.

The fellowship runs for three months, with opportunities to extend the program or transition into a full-time, research-focused role. Fellows work in person alongside researchers and engineers at one of Perplexity's offices and may contribute to research publications and other research outputs.

I came across this opportunity after its priority consideration deadline had passed, but applications are still being accepted on a rolling basis until positions are filled. The priority deadline was September 30, 2026, so applications submitted now may not receive priority consideration. The November 2026 and January 2027 cohorts are listed as priority start dates, but it's worth checking the application page for current availability.

Tracks: Predoctoral and Postdoctoral Research Fellow
Duration: 3 months, full-time
Compensation: $180,000 annualized for predoctoral fellows and $220,000 annualized for postdoctoral fellows
Location: In person in San Francisco, Palo Alto, or New York City
Application status: Accepting applications on a rolling basis until positions are filled
Priority consideration deadline: September 30, 2026 (passed)
Visa support: Available depending on individual circumstances

Note: This is a short-term employment program, not a conventional internship. Students whose requested start date is before graduation should carefully review the program's leave-of-absence requirements.

Fellowship overview | Application

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

Harness Engineering 101: A Practical Guide to Building AI Agents

Who it's for: Developers who want to understand how AI agents work, particularly those interested in LLM APIs, tool use, context management, and building coding agents.

What it's about:

Harness Engineering 101 is a free online book by Isuru Wijesiri that explores how to build AI agents from first principles.

It starts with a simple idea: an LLM provides the reasoning, while the harness is the program that manages the conversation, tools, and execution around it.

The book develops this idea across 16 chapters, using a small Python agent of approximately 300 lines as its foundation. Rather than treating agent frameworks as black boxes, it explains the problems that arise when building an agent and the changes needed to address them.

The chapters cover topics such as the agent loop, tool calling, caching, context-window management, subagents, background work, debugging, guardrails, and retrieval-augmented generation (RAG). There are also sections on coding agents, MCP, skills, hooks, model routing, worktree isolation, retries, rate limits, and streaming.

Each chapter follows a practical structure: the problem being addressed, the minimal change needed to solve it, and the trade-offs that come with that change. The book also includes runnable Python code and diagrams to help explain the concepts.

If you're trying to understand what happens behind an AI agent framework, or want to build a basic agent yourself before working with more complex abstractions, this could be a useful resource to explore.

Format: Free online book with source code on GitHub
Language: Python
Topics: AI agents, LLMs, tool calling, context management, MCP, RAG, coding agents
Version: 1.1, October 2026

Read the book | GitHub repository

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

We've had some really wonderful community finds over the last few weeks, and I honestly can't thank everyone enough for taking the time to find these opportunities, look through them, and send them my way. I know that takes time, and I appreciate every person who thinks of the Radar when they come across something they feel others might find useful 💙

This week, we weren't able to include a community find, but that's completely okay. Hopefully we'll have more wonderful finds from the community as we keep going.

So if you come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, please do send it my way. I love seeing what you find, and it helps make this series better for everyone.

If I feature something you shared in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

🐾 DEV Pet Corner #2: Show Us Your Pets!

Hemapriya Kanagala — Sun, 18 Oct 2026 13:48:11 +0000

TL;DR

DEV Pet Corner is back for another Sunday! It's a weekly series where we take a little break from tech and share our love for animals.

Share pictures, stories, and memories of your pets, animals you've met, or adorable animals you follow online. Share as many pictures and links as you'd like! You don't need to own a pet to join in. If animals make you smile, you're welcome here ❤️

❤️ Thank you for the love
🐶 Show us your animal friends!
🐾 See you next Sunday!

❤️ Thank you for the love!

Hey everyone 👋 It's Sunday again, and DEV Pet Corner is back for its second edition!

First of all, thank you so much to everyone who joined the first edition ❤️ I got to see so many cuties, and I loved going through your comments and meeting all your pets. Some of you also shared memories of pets you've lost, and it meant a lot that you felt comfortable sharing those memories here.

I started this series as a little Sunday break from our usual tech discussions, and seeing everyone join in made me so happy! Getting to spend some time looking at all these adorable animals before getting back to work made my Sunday a little more special.

Thank you for taking the time to share your pics and a little piece of your lives with us. I honestly didn't expect the first edition to turn out this lovely, and I'm so happy I started this!

If you missed meeting all the cuties from the first edition, you can still check it out here -> DEV Pet Corner #1

Now, let's do this again 😄

🐶 Show us your animal friends!

I'd love to see your pets, whether they're the ones waiting for you at home, the ones you grew up with, or the ones who have a special place in your heart.

If it involves an animal you love or something that makes you smile, I'd love to see it 😃

Share pictures, funny moments, sweet memories, little stories, or links to your favorite animal posts. Share as many pictures as you'd like. There's no limit!

If you're sharing someone else's content, please link to the original post and respect the creator's wishes about sharing their work.

🐾 See you next Sunday!

I'm so happy I finally started this little series, and your response to the first edition made me even happier that I did ❤️

As always, let's keep the comments friendly and welcoming to everyone. I'm looking forward to seeing your animal friends and enjoying all the cuteness with you!

Thank you for being here. See you next Sunday for DEV Pet Corner #3!

Dev Opportunity Radar #21: $500K Y Combinator Funding, Bolt Builders Week, and AI Safety Research Residency

Hemapriya Kanagala — Fri, 16 Oct 2026 17:40:16 +0000

TL;DR

Welcome back to Dev Opportunity Radar!

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the MATS Residency for AI safety research, Bolt Builders Week, and applications for Y Combinator's Winter 2027 Batch. I'm also sharing roadmap.sh, a resource with interactive roadmaps and learning materials to help developers figure out what to learn next.

If you're new to the series, you can browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, joined a community, completed a program, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section, with your permission.

And if you've come across an opportunity, resource, community, program, or event that deserves more attention, feel free to share it in the comments.

If I feature one of your Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 MATS Residency: AI Safety Research
- 📌 Bolt Builders Week
- 📌 Y Combinator Winter 2027 Batch
📚 Resources Worth Checking Out
- roadmap.sh: Developer Roadmaps and Learning Resources
🌟 Community Finds
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline
MATS Residency: AI Safety Research	MATS	Research Residency	Hybrid (London, Berkeley, Washington, DC)	October 31, 2026 (AoE)
Bolt Builders Week	Bolt	Build Event & Business Contest	Online	October 20, 2026 (Contest)
Y Combinator Winter 2027 Batch	Y Combinator	Startup Accelerator & Funding	In-person (San Francisco, USA)	November 2, 2026

Resource Highlight

Resource	Why Check It Out
roadmap.sh	Community-driven, interactive roadmaps and learning resources covering developer roles and technologies, along with best practices and questions to test your knowledge.

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications.

So I keep this section to the two most recent editions as a quick reminder of opportunities you may have missed. Older opportunities are still available on the Dev Opportunity Radar website, where you can browse the full collection and use the filters to find something specific.

Opportunity	Organization	Type	Format	Deadline	Featured In
Pear Prime 2027	Pear	Early-Career Program	Hybrid	January 4, 2027	Edition #19
Hacktoberfest 2026	MLH × DEV	Open Source Event	Online + In-person	October 2026	Edition #19
PayPal AI Hackathon	PayPal	AI Hackathon	Online (Worldwide)	November 13, 2026	Edition #20
ARC Prize 2026	ARC Prize	AI Competition & Research	Online / Kaggle	November 2, 2026 (Papers: November 8)	Edition #20
Coinbase Strategy, Execution & Analytics Bounty Challenge	Coinbase	Bounty Challenge	Online (US and Canada, excluding Quebec)	October 18, 2026	Edition #20

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 MATS Residency: AI Safety Research

Who it's for: Researchers, academics, engineers, and professionals from AI, machine learning, policy, governance, and related fields who want to work on AI safety research.

What it's about:

The MATS Residency is a full-time research program focused on reducing catastrophic risks from advanced AI systems. Residents pursue their own research agendas over 6 to 24 months, with access to research advisors, computing resources, and the wider MATS research community.

The program has three tracks for different research goals:

Bridge: For experienced researchers transitioning into AI safety who want to develop a body of relevant research and receive career development and placement support.
Anchor: For researchers looking for a longer-term role to pursue an independent AI safety research agenda.
Incubation: For researchers exploring new research directions or developing AI safety research organizations.

Residents receive $155,000 to $285,000 per year (gross), depending on factors such as experience, circumstances, and office location, along with benefits and research computing resources. The program is based in London, Berkeley, and Washington, DC, with a hybrid arrangement requiring residents to work in person at one of the offices for at least a third of their residency.

The program is open to researchers from a range of backgrounds. Applicants are assessed on the quality and seriousness of their proposed work, with track-specific requirements applying. The residency is full-time, although certain academics on sabbatical may be considered for part-time participation.

Format: Full-time, hybrid research residency
Duration: 6–24 months, depending on track
Locations: London, Berkeley, or Washington, DC
Compensation: $155,000–$285,000 per year, gross, plus benefits
Application Window: August 25–October 31, 2026
Intake: Winter 2027

Residency overview | Track details | Apply

📌 Bolt Builders Week

Who it's for: Builders, founders, entrepreneurs, and anyone interested in creating or growing a business using Bolt.

What it's about:

Bolt Builders Week runs from October 17 to 24, 2026, and includes a free building weekend, live sessions, and a business-building contest.

During the building weekend on October 17 and 18, all Bolt users can access GLM 5.3 Flash through the model picker without changing their plan or redeeming a code. Usage is provided in six-hour blocks, with paid plans receiving five times the free allowance per block. Standard rate limits apply, and unused allowance does not carry over.

Live sessions run from October 20 to 23, covering topics such as SEO, customer onboarding and retention, design, and branding. The schedule also includes a Q&A with Bolt co-founders Eric Simons and Albert Pai.

The contest invites participants to build a business idea into a product or use Bolt to develop an existing business. Projects must be built primarily in Bolt, and submissions must include a publicly accessible live link and a Bolt project link. A demo video is optional.

The prizes are $10,000 for first place, $5,000 for second place, and $2,500 for third place. The first-place prize also includes a one-on-one mentoring session with Bolt co-founder and CEO Eric Simons. Entries are judged equally on use of Bolt, the idea, implementation and design, and potential impact.

Format: Online building weekend, live sessions, and contest
Dates: October 17–24, 2026
Deadline: October 20, 2026
Prizes: $10,000, $5,000, and $2,500
Eligibility: Contest entrants must be at least 18 and meet the official rules

Explore Bolt Builders Week

📌 Y Combinator Winter 2027 Batch

Who it's for: Founders who are building a startup or have an idea they want to develop into a business.

What it's about: Y Combinator (YC) is accepting applications for its Winter 2027 startup funding batch. The program runs from January to March 2027 and takes place in person in San Francisco.

YC invests $500,000 in each accepted company. During the three-month program, founders work with a dedicated YC partner, attend group sessions and meetups, and get support with developing their products and raising further investment.

Startups at different stages can apply, including those that are still at the idea stage. Applications are open to startups from anywhere in the world. However, founders should be prepared to work full-time on their company during the batch, which takes place in San Francisco.

If the upcoming batch doesn't fit your timeline, YC also accepts applications for future batches through its Early Decision option.

Format: In-person startup accelerator and funding program
Batch Dates: January to March 2027
Location: San Francisco, California, USA
Investment: $500,000 per accepted company
Application Deadline: November 2, 2026, at 8 PM PT

Decision Date: Applicants who apply by the deadline will receive a decision by December 11, 2026.

Apply to Y Combinator | Learn about Early Decision

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

roadmap.sh: Developer Roadmaps and Learning Resources

Who it's for: Developers who want to learn a new technology, prepare for a role, or figure out what skills to learn next.

What it's about: roadmap.sh is a community-driven platform with roadmaps, study plans, and learning resources for different developer roles and technologies.

The roadmaps are interactive, so you can click on individual topics to explore them and find resources to learn more. There are paths for frontend, backend, DevOps, cybersecurity, AI engineering, data science, cloud, and many other areas.

Beyond roadmaps, the platform also has best practices for topics like API security and code review, along with questions to test your knowledge of technologies such as JavaScript, React, and Node.js.

You can browse the roadmaps and use them to plan your learning. The project also accepts community contributions to its roadmaps.

Format: Online learning resource
Cost: Free to access the roadmaps and resources

Explore roadmap.sh

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

This week, we weren't able to include a community find, but that's completely okay. Hopefully we'll have more wonderful finds from the community as we keep going.

If I feature something you shared in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

🐾 DEV Pet Corner #1: Show Us Your Pets!

Hemapriya Kanagala — Sun, 11 Oct 2026 13:52:25 +0000

TL;DR

Welcome to DEV Pet Corner, a new weekly series where we take a little break from tech and share our love for animals!

Every Sunday, let's share pictures, stories, and memories of our pets, animals we've met, and even the adorable animals we love following on Instagram or elsewhere online. Share as many pictures and links as you'd like! You don't need to own a pet to join in. If animals make you smile, you're welcome here ❤️

📚 Table of Contents

❤️ Why I'm starting this
🐶 Show us your animal friends!
🐾 See you next Sunday!

❤️ Why I'm starting this

I've wanted to start this series for the longest time, but somehow, I kept putting it off. Today, I'm finally doing it!

If there's one thing you should know about me, it's that I LOVE PETS.

My family had a dog when I was very young, and although we haven't had a pet since, I've always had a soft spot for animals.

When I was doing my master's at the University of Oklahoma, I used to take pictures of dogs I came across on campus. Being away from home, those little encounters often made my day. Sometimes, a quick moment with a dog between classes was enough to brighten an otherwise ordinary day. I still have those pictures, and looking back at them brings back some lovely memories ❤️

If you've read my personal reflection posts, you probably know that I often talk about how our health and well-being come first. We're so busy studying, building careers, and working towards our goals that we sometimes forget to slow down and enjoy life along the way.

That's what inspired this series. We spend so much time talking about tech here on DEV, and I thought it would be lovely to have a little Sunday tradition where we can share something different, get to know each other beyond our work, and enjoy the simple happiness that animals bring us 💛

🐶 Show us your animal friends!

Now, let's get to the fun part!

I'd love to see your pets, whether they're the ones waiting for you at home, the ones you grew up with, or the ones who have a special place in your heart.

If it involves an animal you love or something that makes you smile, I'd love to see it!

Share pictures, funny moments, sweet memories, little stories, or links to your favorite animal posts. Share as many pictures as you'd like. There's no limit!

If you're sharing someone else's content, please link to the original post and respect the creator's wishes about sharing their work.

Don't have anything to share this time? No worries! Just enjoy all the adorable animals in the comments, and join us whenever you'd like.

I'll share a picture of a very cute dog I came across at the University of Oklahoma in the comments. I'm happy to finally share a little memory from my time there!

🐾 See you next Sunday!

DEV Pet Corner will be here every Sunday 😃

I've wanted to start this series for a long time, and I'm so happy I've finally decided to give it a try.

I'm not doing this for views, likes, or comments. I simply love animals, and I thought it would be lovely to have something a little different to look forward to each week.

Even if nobody joins in this first edition, I'll still be back next Sunday, and the Sunday after that. Sometimes, an idea is worth trying simply because it means something to you.

If you have any ideas for future editions, I'd love to hear them!

One little note: I know you all are wonderful and kind, but since this is a little off-topic from our usual tech posts, I just wanted to mention that let's keep the comments friendly, positive, and welcoming to everyone.

I'll be checking the comments, and I'm really looking forward to seeing your animal friends, reading your stories, and enjoying all the cuteness with you!

Thank you for being here for the very first DEV Pet Corner! See you next Sunday!

Dev Opportunity Radar #20: $2M AI Reasoning Challenge, PayPal AI Hackathon, and Coinbase Analytics Challenge

Hemapriya Kanagala — Fri, 09 Oct 2026 14:55:22 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the PayPal AI Hackathon, ARC Prize 2026, and Coinbase's Strategy, Execution & Analytics Bounty Challenge, along with Build Your Own X as a resource for learning how different technologies work by building them from scratch.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, joined a community, completed a program, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, or event that deserves more attention, feel free to share it in the comments.

If I feature one of your Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 PayPal AI Hackathon
- 📌 ARC Prize 2026
- 📌 Coinbase Strategy, Execution and Analytics Bounty Challenge
📚 Resources Worth Checking Out
- Build Your Own X
🌟 Community Finds
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline
PayPal AI Hackathon	PayPal	AI Hackathon	Online (Worldwide)	November 13, 2026
ARC Prize 2026	ARC Prize	AI Competition & Research	Online / Kaggle	November 2, 2026 (Papers: November 8)
Coinbase Strategy, Execution & Analytics Bounty Challenge	Coinbase	Bounty Challenge	Online (US and Canada, excluding Quebec)	October 18, 2026

Resource Highlight

Resource	Why Check It Out
Build Your Own X	A collection of step-by-step tutorials for recreating technologies from scratch, covering topics such as databases, operating systems, neural networks, AI models, programming languages, and web servers.

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications.

So I keep this section to the two most recent editions as a quick reminder of opportunities you may have missed. Older opportunities are still available on the Dev Opportunity Radar website, where you can browse the full collection and use the filters to find something specific.

Opportunity	Organization	Type	Format	Deadline	Featured In
Nebius x NVIDIA Global AI Hackathon	Nebius × NVIDIA	Hackathon	Online	October 30	Edition #18
Pear Prime 2027	Pear	Early-Career Program	Hybrid	January 4, 2027	Edition #19
Graph Hacks: Context for AI Agents	WeMakeDevs × FalkorDB	AI Hackathon	Online	October 18	Edition #19
Hacktoberfest 2026	MLH × DEV	Open Source Event	Online + In-person	October 2026	Edition #19

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 PayPal AI Hackathon

Who it's for: Developers, designers, founders, students, and anyone interested in building with AI and payment technologies.

What it's about:

The PayPal AI Hackathon is a global online hackathon where participants build an application that meaningfully integrates PayPal's developer platform and an AI tool, model, or platform of their choice.

There are no fixed problem statements or required tracks. Projects can include AI agents, payment experiences, business tools, applications, or automations. Participants can work individually or in teams, and they can start a new project or make meaningful progress on an existing one during the hackathon.

A working prototype or proof of concept is required. Submissions must include a public code repository with a license, documentation, a functional demo, details of the AI tools used, and a demo video under three minutes. Static mockups and non-functional demos do not meet the requirements.

The total prize pool is listed as $67,500+, with awards for the overall winners and categories such as creativity, impact, agentic commerce, and the use of PayPal and sponsor tools. Some awards include cash, while others offer credits or subscriptions.

Format: Online, worldwide

Dates: Deadline November 13, 2026, at 1:30 AM IST

Team Size: Individual or team

Prize Pool: $67,500+ in cash and other prizes

Requirement: Meaningful use of both PayPal and AI

Explore the hackathon

📌 ARC Prize 2026

Who it's for: AI researchers, developers, and anyone interested in reasoning systems, AI agents, and open-source AI research.

What it's about:

ARC Prize 2026 is a competition focused on advancing AI reasoning, with $2 million in total prizes across three tracks. Participants can work on static reasoning tasks, interactive AI agents, or research papers exploring approaches to improving performance on the ARC-AGI benchmark.

There are three tracks:

ARC-AGI-2: Build systems that solve static reasoning tasks by predicting the correct outputs for visual grid-based problems.
ARC-AGI-3: Build AI agents that explore unfamiliar environments, learn how they work, set goals, and plan actions without being given explicit instructions.
Paper Prize: Submit a research paper presenting an approach to improving ARC-AGI performance, supported by a corresponding Kaggle code submission.

The prize pools differ by track. ARC-AGI-2 has $700,000 available, ARC-AGI-3 has $850,000, and the Paper Prize has $450,000. Awards include prizes for high-scoring solutions, research contributions, and progress toward specific benchmarks.

One important requirement is that participants must open-source their solutions to be eligible for prizes. Submissions are made through the designated Kaggle competitions, and internet access is unavailable during evaluation, so API-based systems such as GPT or Claude cannot be used during that stage.

Format: Online, through Kaggle

Competition Start: March 25, 2026

Submission Deadline: November 2, 2026

Paper Deadline: November 8, 2026

Results: December 4, 2026

Total Prize Pool: $2 million across three tracks

Requirement: Reproducible, open-source solutions for prize eligibility

Explore the competition

Track details:

📌 Coinbase Strategy, Execution and Analytics Bounty Challenge

Who it's for: People interested in analytics engineering, data modeling, SQL, Python, and data quality.

What it's about:

Coinbase's Strategy, Execution & Analytics team is running a bounty challenge built around a fictional scenario involving an Analytics Engineer intern supporting the Growth team.

Participants must complete four exercises covering data modeling for an onboarding funnel, investigating inconsistencies across signup reports using Python, reviewing an inherited SQL pipeline, and communicating a metric discrepancy to a product manager.

The challenge focuses on technical skills as well as decision-making, data validation, and explaining data issues to stakeholders. No live dataset is provided, so participants are expected to explain their assumptions and describe their solutions using code, schemas, or written responses where appropriate.

Submissions must include a current resume in PDF format, responses to all four exercises within a combined limit of 1,500 words (excluding code blocks), and a brief disclosure explaining how AI tools were used, if applicable.

The top 20 participants are eligible to receive $250 in USDC and an exploratory call with a Coinbase Emerging Talent recruiter. Participation does not count as a job application or guarantee an interview or internship.

Format: Online bounty challenge

Challenge Window: September 25 to October 18, 2026

Deadline: October 18, 2026, at 11:59 PM Pacific Time

Prize: $250 in USDC per selected participant, plus a recruiter call

Number of Recipients: 20

Eligibility: Legal residents of the United States or Canada, excluding Quebec, aged 18 or older, with a verified Coinbase account

Challenge details and submission

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

Build Your Own X

Who it's for: Developers and learners who want to understand how programming concepts, tools, and technologies work by building them from scratch.

What it's about:

Build Your Own X is a GitHub repository maintained by CodeCrafters that collects step-by-step tutorials for recreating different technologies. Topics include databases, neural networks, programming languages, operating systems, web servers, Git, Docker, AI models, and more.

The repository links to tutorials in different programming languages, covering projects ranging from building a simple neural network to implementing a database or creating a programming language. Readers can browse the categories and choose a project based on what they want to learn.

It's a collection of external tutorials rather than a single structured course, so the languages, prerequisites, and difficulty levels vary between projects.

Format: Free, open-source collection of tutorials

Topics: Software Engineering, Systems Programming, AI/ML, Databases, Networking, and more

Languages: Varies by tutorial

Cost: Free

Explore the repository

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

This week, we weren't able to include a community find, but that's completely okay. Hopefully we'll have more wonderful finds from the community as we keep going.

If I feature something you shared in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

The Parts of a Job Search We Don't See

Hemapriya Kanagala — Mon, 05 Oct 2026 15:28:28 +0000

TL;DR

We usually see a job search as applying, interviewing, and hopefully getting an offer. But there is a lot happening around those steps that we rarely get to see.

I recently read Notes on the Industry Job Search by Alisa Liu, a researcher at OpenAI on the pretraining team, where she shares what her own industry search looked like after a six-year NLP PhD. Reading her experience made some parts of the job search much clearer to me, especially the things that are easy to underestimate: timing, interview stamina, preparing differently for every process, getting help from people around you, negotiation, and the emotional side of it all.

These are the things I took away from her experience.

The full article is linked in the References section at the end. Please read it if you want the complete picture, because there is much more detail in her post than I can cover here.

And if you have been through a job search yourself, I’d love to hear about your experience too. What did you learn along the way? What helped you? What do you wish you had known earlier? Feel free to share it in the comments so others can learn from your experience too ❤️

What I took away
- The job search is much bigger than the interview
- Timing your interviews is not as simple as it sounds
- Your first interview may have been prepared for long before you knew you needed it
- Having the experience to get the interview and having the skills to pass it are two different things
- "Technical preparation" is not one giant category
- You can know something and still not be ready to demonstrate it
- The interviewer may be tired too
- Your preparation should change depending on the opportunity
- Interview stamina is a real thing
- Writing down what happened after an interview can become part of your preparation
- The studying can change you beyond the interview
- Behavioral interviews can be harder precisely because they look easy
- Getting an offer can create another kind of uncertainty
- An initial offer isn't necessarily the final conversation
- A lot of career decisions are made without enough information
- The emotional side is not a side note
- There is a difference between preparing for the next chapter and missing the current one
Please take care of yourself
I'd love to hear from you
References
Let's Stay Connected

What I took away

I don't want this to be a summary of Alisa's article. Her original post already does that much better than I could.

Instead, I wanted to share the parts that made me stop and think, especially the things that I don't usually see talked about when people talk about job searching.

Some of these are practical, like how you prepare, how you approach different interviews, and what happens after an offer. Others are much more personal, like dealing with uncertainty, comparison, exhaustion, and remembering to take care of yourself through the process.

So, here are the things I took away from her experience.

1. The job search is much bigger than the interview

When someone says, "I'm job searching," I think we often picture a very small part of what is actually happening.

Apply, hear back, interview, get an offer.

But Alisa's numbers gave me a much better sense of what can actually be behind that process. She interviewed at 11 companies across 57 interviews. On top of that, she had 46 additional recruiter calls, 16 post-offer conversations, and many informal networking conversations.

That is a lot of work that you would never know about if all you saw was the final sentence:

"I interviewed with 11 companies."

There is deciding when to start each process, preparing for each interview, talking to recruiters, keeping track of different companies, recovering between interviews, making decisions with incomplete information, and then potentially doing all of that again after an offer.

The interview is only one visible part of the job search.

And I think seeing that more clearly changes how we look at our own progress too. Sometimes what feels like "nothing is happening" may actually involve a lot of work that simply isn't visible from the outside.

2. Timing your interviews is not as simple as it sounds

This was one of the less obvious things I took from her article. There is common advice that says you should use a few companies as practice and then try to time the companies you care about so that the offers arrive around the same time. That sounds very neat when written as a strategy.

Real life is obviously messier. Alisa points out that companies can have different hiring timelines, different amounts of headcount, different teams hiring at different times, and different levels of urgency.

So sometimes the timing of your process is influenced by things that have nothing to do with how prepared you are. That was a useful reminder for me because job searches can sometimes make us believe every outcome is a direct reflection of our own decisions.

It isn't always. There are things happening on the other side of the table too.

And even offer deadlines can have more flexibility than candidates might assume, while some offers can have much tighter deadlines. Knowing that these things exist can help you ask better questions instead of simply accepting every timeline as fixed.

3. Your first interview may have been prepared for long before you knew you needed it

This was another part I found interesting. Alisa talks about doing good work during the PhD, making friends, collaborating, going to conferences, and building relationships. Then, when the job search started, some of those connections could become helpful in getting that first interview.

That made me think about networking differently. Sometimes we treat networking as something we have to suddenly start doing when we need a job. But a relationship built through a project or collaboration years ago can become useful much later.

She also mentions reconnecting with people she had not spoken to in years. That is something I don't think people talk about enough. You don't necessarily have to know hundreds of people. Sometimes it is simply about not being afraid to reach out to someone you already know, even if it has been a long time.

And there is another part here that I loved. She also says networking does not come easily to everyone and that people should take care of their own energy and comfort levels. So this isn't about forcing yourself to become the most social person in the room.

Build relationships in ways that feel possible for you.

4. Having the experience to get the interview and having the skills to pass it are two different things

This was a very useful thing to notice in her experience. Alisa says that technical skills and knowledge were evaluated much more heavily during her interviews than research experience, while her research background probably helped get her the interviews in the first place.

That is an important difference. Your background can make someone interested enough to talk to you. But once you are in the interview, you still have to demonstrate what you can actually do.

Those are two different stages. And I think that applies beyond research roles too.

A strong resume can open a door.

It doesn't replace being able to walk through it.

5. "Technical preparation" is not one giant category

Alisa's interviews included ML coding, general coding, technical discussions, research discussions, behavioral interviews, math, and a job talk.

What I found useful here is not just the list itself. It is what the list tells us.

Being technically strong does not necessarily mean being prepared for every way someone can ask you to demonstrate that strength.

One interview might ask you to implement something. Another might ask you to explain why a particular approach works. Another might ask you to design an experiment. Another might ask you to discuss your own research.

Another might ask you to solve a math problem on paper. Another might ask you about a situation where you disagreed with someone.

These are very different things.

So preparing for "the interview" is not really enough. You need to prepare for the kind of thinking the interview is actually asking you to demonstrate.

6. You can know something and still not be ready to demonstrate it

This is one of my favorite practical points from the article. Alisa describes taking notes, drawing diagrams, practicing problems, implementing things from scratch, and spending entire days making sure she understood fundamental ML concepts deeply.

But she also specifically recommends practicing coding with AI assistance completely off when preparing for interviews. That feels especially relevant now.

In everyday work, using AI can be completely normal. We can ask for an explanation, check an approach, debug something, or get unstuck. But if the interview doesn't allow those tools, practicing with them every time can give you a very different picture of your readiness.

You might understand a concept very well while still finding it difficult to produce the solution yourself under pressure.

So there is a difference between:

"I understand this when I have help."

and

"I can reason through this myself when someone puts the problem in front of me."

Both are useful skills. But they are not the same skill.

7. The interviewer may be tired too

This is a small point in her article, but I thought it was surprisingly practical. When talking about her research pitch, Alisa mentions tailoring it to the role and making it easier for interviewers to see why her background was relevant.

Part of her reasoning was simply that interviewers are tired. That sounds obvious once you hear it, but it changes how you think about communication.

You may know your work extremely well. You may have spent years on it. But the person listening to you may have a long day of interviews, meetings, and other responsibilities.

Making the relevant part easy to understand isn't about making your work less sophisticated. It is about helping another person understand it.

That is a useful communication skill far beyond interviews.

8. Your preparation should change depending on the opportunity

Alisa describes looking at the interview description, topics the company cares about, information from recruiters, and what she knew about the company to understand what she should focus on.

That sounds simple, but it changes the question you ask yourself.

Instead of:

"What should I study for interviews?"

it can be:

"What is this particular interview actually trying to find out about me?"

That second question is much more useful.

You might need more coding practice for one role. You might need to review fundamentals for another. You might need to prepare your research story. You might need to spend more time on behavioral examples. You might need to brush up on math.

There is value in having strong foundations, but there is also value in not spending your limited preparation time equally on everything.

9. Interview stamina is a real thing

This is something I rarely see discussed when people give interview advice. Alisa talks about using some companies as practice, but also warns that stamina is finite.

That matters when you are interviewing with several companies at once. You are not just solving problems. You are switching between topics, remembering different processes, talking to recruiters, preparing presentations, studying, and making decisions. And then you are expected to show up and think clearly again.

She also shares that she once went into a technical interview after sleeping only 2 hours because she had been cramming. The knowledge she had tried to put into her head at the last minute did not end up helping her.

Instead, she spent time making a simple mistake because she was exhausted.

That made one thing very clear to me.

More preparation is not always better preparation.

Sometimes the most useful thing you can do before an interview is sleep.

10. Writing down what happened after an interview can become part of your preparation

Another small detail from her process that I thought was useful was recording notes after interviews. That makes sense for a reason beyond remembering the questions.

If you are doing several interviews, you can quickly forget what you struggled with, what you explained well, what you wish you had answered differently, or what you learned.

Writing those things down turns one interview into information you can use for the next one.

Instead of every interview being an isolated event, you start building your own feedback loop.

And that is especially useful when the process stretches over weeks or months.

11. The studying can change you beyond the interview

This was probably one of the most meaningful parts of the article for me. Alisa says that studying for interviews gave her a wider breadth of knowledge, improved her confidence in technical conversations, and even helped her become more effective in her research.

That is a much bigger benefit than simply passing an interview.

Sometimes we think of interview preparation as temporary studying. Learn this because they might ask. Practice this because it might come up. Memorize this because it is on the list.

But deep learning does not disappear when the interview ends. Understanding something properly can change the questions you ask, the problems you notice, the conversations you can participate in, and the ideas you can come up with.

That made me think about interview preparation less as studying only for a test and more as a chance to find gaps in my own understanding.

12. Behavioral interviews can be harder precisely because they look easy

Alisa shares that she failed her first behavioral interview because she had not prepared for what she thought would be very simple questions. I can see how easy it is to fall into that trap.

You might think:

"I know myself. Of course I can answer a question about myself."

Then someone asks you for a specific example from your experience, and suddenly you are trying to remember what happened three years ago while also figuring out how to answer the actual question.

Her solution was to think through memorable stories from her PhD and map them to common behavioral questions so that she could retrieve the right example more easily.

That doesn't mean memorizing fake answers. It means knowing your own experiences well enough that you can talk about them clearly.

13. Getting an offer can create another kind of uncertainty

I think this is one of the easiest parts of a job search to overlook. We usually imagine the offer as the finish line.

But in her experience, it opened another stage. There were conversations with future teammates and managers, recruiter calls, visits, more communication, comparing options, and negotiation.

And negotiation was not something she felt naturally prepared for. She says that unlike interviews, it was not something she could simply study her way through.

She was also very honest about the imbalance of experience. Recruiters negotiate these things regularly. Most candidates don't.

That is why she leaned on friends for advice and for more data points.

That was a useful reminder that sometimes preparation means recognizing where you don't have enough information yourself and finding people who do.

14. An initial offer isn't necessarily the final conversation

This is another point that can be easy to miss if you have never been through the process. Alisa says recruiters sometimes explicitly told her that they did not expect candidates to simply accept the first offer.

She also talks about how putting effort into negotiation for a few weeks can have a very meaningful financial impact.

I don't think the takeaway is "always negotiate aggressively." It is more basic than that.

If you receive an offer, you are allowed to ask questions. You are allowed to understand what is flexible. You are allowed to think about what matters to you.

And you don't have to assume that the first number you see is automatically the only possible outcome.

15. A lot of career decisions are made without enough information

This may be one of the hardest things to understand until you experience it yourself. Alisa talks about making decisions with incomplete information and how even small choices, such as who to contact and when, can feel like they have much bigger consequences than they should.

There often isn't a perfect answer. You don't know exactly what a team will be like six months from now. You don't know whether another opportunity will appear. You don't know whether waiting will help or hurt. You don't know why one company moves quickly while another goes quiet.

And sometimes you have to make a decision anyway.

I think that is an important part of a job search to accept. You can prepare carefully without being able to control everything.

16. The emotional side is not a side note

This part of her article is very honest. She talks about comparing herself with peers, dealing with people's opinions about where she should go, managing a huge number of decisions, and feeling stressed and miserable for months.

That is very different from the usual career advice where everything is reduced to a list of things you should optimize.

Real people don't experience a job search like a spreadsheet.

You can prepare properly and still be scared. You can get an interview and still wonder whether you're good enough. You can receive an offer and still not know what you want.

You can watch your friends move ahead and feel like you're falling behind even when you aren't. And sometimes the uncertainty itself is exhausting.

I think there is value in saying that out loud because someone going through it might otherwise assume they are the only person finding it difficult.

17. There is a difference between preparing for the next chapter and missing the current one

I really loved where Alisa takes the article at the end. After writing about the stress of the job search, she talks about how sad she was to leave her PhD and how special that period had been.

There is something very human about that.

We are constantly preparing for whatever comes next. As students, we prepare for jobs. When we start working, we prepare for the next role. Then the next promotion. Then the next move.

There is always another thing waiting.

Preparing for the future is important. But it should not mean that the present becomes something we are only trying to get through.

If you are learning something right now, there can still be joy in learning it. If you are working on a project, there can still be joy in building it. If you are studying, you don't have to turn every minute into career preparation.

You are allowed to enjoy the chapter you are currently living too.

Please take care of yourself

I wanted to make this a separate section because this is something I tell people all the time, and I really mean it.

If you are currently searching for a job, please take care of yourself with the utmost love.

I know the job market is hard. I know how much pressure can come with searching. I know rejection can hurt. I know seeing other people get offers can make you question yourself. I know how easy it is to feel like you should always be doing one more application, one more problem, one more certification, one more interview, one more thing.

But please take care of yourself too.

Your physical health matters. Your mental health matters. Your sleep matters. Your food matters. Your relationships matter. Your time away from the screen matters. Your peace matters.

Because if you are not okay, what are all those other things for?

A job is important. An offer is important. Your career is important. But none of those things are worth losing yourself over.

I don't mean this as a reminder to "balance" everything perfectly. Life does not work that neatly. There will be days when you have more energy and days when you don't. There will be days when you can study for hours and days when even opening your laptop feels difficult.

And that is okay.

You don't have to feel guilty for resting. You don't have to feel guilty for having a life outside your job search. You don't have to feel like every hour that you are not applying or studying is an hour wasted.

You are a person before you are a job seeker.

Please don't be cruel to yourself because the market is difficult. Please don't decide that someone else's timeline is proof that you are behind. Please don't measure your entire worth through responses from companies.

Sometimes things simply take longer than we hoped they would.

And I know it can be very hard to believe that everything will be okay when you are in the middle of it. I know there can be bills to worry about, family expectations, loans, uncertainty about the future, and so many other things that make "just stay positive" feel impossible.

So I am not asking you to pretend everything is fine.

I am just asking you to be a little kinder to yourself while you are going through it.

Do what you can. Keep going when you have the energy. Rest when you don't. Ask for help when you need it. Talk to someone you trust. Take a break when you need one.

And please remember that a rejection is not a measurement of your worth. A delayed offer is not a measurement of your potential. Someone else's success is not proof that you are failing.

Your life is so much bigger than this job search.

There will be things to celebrate that have nothing to do with a job offer. There will be people who love you regardless of what is written on your resume. There will be mornings when you wake up and realize that this difficult chapter did eventually pass.

So please keep going, but please don't forget to live while you are going.

You matter in this process too.

And if you are reading this while going through a difficult job search, I really hope you are being gentle with yourself.

Please eat. Please sleep. Please step away from the screen. Please talk to someone. Please let yourself laugh. Please let yourself have a good day without feeling like you should have been applying instead.

You do not have to earn the right to take care of yourself by getting a job first.

And if today was a difficult day, that's okay.

Try again tomorrow.

With the utmost love, I hope you are okay. I hope you are taking care of yourself. And I truly hope something good finds you soon ❤️

I'd love to hear from you

If you have already been through a job search, or you are currently going through one, I would genuinely love to hear from you.

Maybe you have a preparation method that helped you. Maybe there is something you wish someone had told you earlier. Maybe you learned something from a rejection. Maybe you have a networking tip.

Maybe you learned something about interviews, negotiation, choosing between offers, or simply getting through the difficult parts.

And if you are comfortable sharing, please do.

There are people here at very different stages of their careers. There are students, new graduates, early-career developers, experienced professionals, people changing careers, and people looking for their next opportunity.

I'm early in my own career too, so I would genuinely love to learn from people who have already been through some of this.

Maybe we can make the comments section into something useful for someone who finds this post months or years from now.

Share whatever you feel comfortable sharing.

References

Notes on the Industry Job Search by Alisa Liu

This is the original article I read, and I really recommend giving it a read if you have the time.

I have only written about the parts that I personally took away from her experience, but her original post gives a much fuller view of the journey. She goes into her timeline, the different interview formats, how she prepared, how she approached different interview processes, recruiter conversations, negotiation, the emotional side of the search, and the resources she used to study.

There is much more detail in her post than I can reasonably fit here, and reading the full thing gives you the context behind all of these points.

For me, it is a very good read, especially if you are preparing for an industry job search or simply want to understand what the process can look like from the other side.

🤝 Let's Stay Connected

Place	Find me here
GitHub	building things → hemapriya-kanagala
LinkedIn	resources & updates → hemapriya-kanagala
X	random dev thoughts → @KanagalaHema

Dev Opportunity Radar #19: Hacktoberfest, $20K AI Agent Hackathon, and Pear Prime 2027

Hemapriya Kanagala — Fri, 02 Oct 2026 15:16:19 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features Pear Prime 2027, Hacktoberfest 2026, and Graph Hacks: Context for AI Agents, along with How to Scale Your Model as a resource worth checking out.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, joined a community, completed a program, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, or event that deserves more attention, feel free to share it in the comments.

If I feature one of your Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Pear Prime 2027
- 📌 Hacktoberfest 2026
- 📌 Graph Hacks: Context for AI Agents
📚 Resources Worth Checking Out
- How to Scale Your Model
🌟 Community Finds
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline
Pear Prime 2027	Pear	Early-Career Program	Hybrid / Community + Career	January 4, 2027
Hacktoberfest 2026	MLH × DEV	Open Source Event	Online + In-person	October 2026
Graph Hacks: Context for AI Agents	WeMakeDevs × FalkorDB	AI Hackathon	Online (Worldwide)	October 18, 2026

Resource Highlight

Resource	Why Check It Out
How to Scale Your Model	A free online book covering the systems side of scaling language models, including TPUs, GPUs, sharding, parallelism, training, inference, and performance.

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications.

So I keep this section to the two most recent editions as a quick reminder of opportunities you may have missed. Older opportunities are still available on the Dev Opportunity Radar website, where you can browse the full collection and use the filters to find something specific.

Opportunity	Organization	Type	Format	Deadline	Featured In
Nebius x NVIDIA Global AI Hackathon	Nebius × NVIDIA	Hackathon	Online	October 30	Edition #18
Build, Ship, Shape: Amazon Developer Hackathon	Amazon	Hackathon	Online	October 24	Edition #17
Dev3Pack Hackathon	Dev3Pack	Hackathon	Hybrid	October 30	Edition #17

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Pear Prime 2027

Who it's for: CS students, new graduates, and early-career engineers.

What it's about: Pear Prime is a program from Pear for a small cohort of early-career engineers. The program connects participants with founders and technical leaders from startups and technology companies, along with mentorship and career support from the Pear team.

For the 2027 cohort, Pear lists companies including Cerebras, Etched, Perplexity, HeyGen, Baseten, Forus, Vals, and Andera as teams that Prime Engineers can get exposure to.

The program also includes a community of other Prime Engineers, mentorship and career coaching, access to Pear investors for guidance and feedback on ideas, and priority consideration for PearX if members later decide to start a company.

The application is fairly focused on what you've built and where you want to go next. You'll be asked about your most significant project, internship, or research experience, technologies you're strongest in, notable achievements, the type of engineering work you enjoy, and what you're hoping to build or become over the next few years.

There is also a question about something you built or shipped because you wanted to, rather than because it was required by a class, job, or manager. So having personal projects, open-source work, research, or other examples of self-directed work could be useful to have ready when applying.

Format: Hybrid / Community + career opportunities
Program: Pear Prime 2027
Who can apply: CS students, new graduates, and early-career engineers
Application Deadline: January 4, 2027
Focus: Engineering careers, startups, mentorship, community, and building
Application: Resume + project/research experience + technical interests + career goals + other background information

🔗 Learn More | 🔗 Apply

📌 Hacktoberfest 2026

Who it's for: Anyone interested in open source, open-source AI, or building with open-weight models.

What it's about: Hacktoberfest 2026 is a month-long celebration of open source happening throughout October, with both online and in-person activities.

This year's Hacktoberfest is focused on open-source AI and open-weight models, with Fests happening in cities around the world as well as online activities that anyone can join.

If you're participating online, there are livestreams, Global Hack Week, and DEV Challenges, with a new mini-hackathon launching on DEV each week during October.

You can participate online from anywhere, attend an in-person Fest, or do both. And you don't need to be an experienced open-source contributor to take part.

One thing that's different this year is that pull requests and merge requests no longer count toward Hacktoberfest rewards. The event is instead focusing more on learning, building, and participating in the different activities.

Since I'm here on DEV and Hacktoberfest is powered by MLH and DEV, I wanted to give it a small mention here too. If you're already part of the DEV community, you might want to take a look at the DEV Challenges happening throughout October.

Of course, as always, it's completely up to you whether any of the activities are something you'd like to take part in. I just wanted to make sure people here knew it was happening.

Format: Online + In-person | When: October 2026
Cost: Free
Focus: Open source, open-source AI, and open-weight models
Online Activities: DEV Challenges, livestreams, and Global Hack Week
In-Person: 300+ Fests worldwide

🔗 Explore Hacktoberfest 2026
🔗 See DEV Challenges

📌 Graph Hacks: Context for AI Agents

Who it's for: Developers, AI builders, and anyone interested in building AI agents.

What it's about: Graph Hacks: Context for AI Agents is an online hackathon by WeMakeDevs and FalkorDB running from October 15–18, 2026.

The challenge is to build an AI agent or group of agents that uses FalkorDB as the primary graph database. The project should use the graph as part of what the agent does, such as working with connected data, remembering information between sessions, sharing state between agents, or connecting company knowledge.

There are three tracks:

Agents That Act on Connected Data
Agent Memory and Coordination
Company Brain

You can participate in one, two, or all three tracks. You can join solo or with a team of up to four, and you don't need previous graph database experience.

One thing to keep in mind is that FalkorDB must be a central part of the project, and the main FalkorDB implementation needs to be new work completed during the hackathon. Coding assistants are allowed.

There is a $20,000 prize pool, with prizes for each track, along with prizes for the best project blog posts and the top 50 submissions.

Format: Online | Worldwide | Dates: October 15–18, 2026
Team Size: Solo or up to 4
Cost: Free
Tracks: 3
Graph Experience: Not required
AI Coding Assistants: Allowed
Deadline: October 18, 2026

🔗 Learn More & Register

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

How to Scale Your Model

If you're interested in understanding what happens when language models have to run across large amounts of hardware, How to Scale Your Model is a free online book from researchers at Google DeepMind and other institutions that takes a systems view of scaling LLMs on TPUs and GPUs.

The book looks at how hardware and models interact, including compute, memory, communication, model parallelism, training, inference, and scaling across multiple devices.

It starts with topics like roofline analysis, how TPUs work, and sharded matrix multiplication, then moves into Transformer math, training and inference parallelism, practical LLaMA 3 tutorials, profiling TPU code, JAX, and a bonus section on GPUs.

Some of the topics covered include data, tensor, pipeline, and expert parallelism, along with techniques such as model sharding, rematerialization, host offload, and gradient accumulation.

The resource is fairly technical. It assumes that you already understand the basics of LLMs and the Transformer architecture, and ideally have some familiarity with JAX. So this is probably more useful if you're already comfortable with the fundamentals and want to understand what happens underneath the model when it needs to scale.

The book is organized into 12 chapters, and the authors mention that you don't have to read everything in order. The first few chapters introduce the underlying concepts, while later sections apply them to real models and practical systems.

Format: Free online book
Focus: LLM systems, TPUs, GPUs, Transformers, training, inference, parallelism, and performance
Level: Advanced / technical
Prerequisites: Basic LLM and Transformer knowledge; JAX familiarity is helpful
Topics: Roofline analysis, TPU architecture, sharding, Transformer math, training, inference, profiling, JAX, GPUs
Authors: Researchers and engineers from Google DeepMind, Stanford, and other organizations

🔗 Read How to Scale Your Model

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

This week, we weren't able to include a community find, but that's completely okay. Hopefully we'll have more wonderful finds from the community as we keep going.

If I feature something you shared in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

I Built My First AI Agent With AWS AgentCore, and the Hardest Part Wasn't the AI

Hemapriya Kanagala — Tue, 29 Sep 2026 15:44:51 +0000

TL;DR

I recently completed another project from Udacity's Future AWS Agent Engineer Nanodegree Program, which I was able to take through the AWS AI & ML Scholarship.

This was my second project working with AI agent systems, but it was my first time building one with Amazon Bedrock AgentCore and the Strands SDK. The goal was to build a customer support AI agent that could track orders, process refunds, answer product and policy questions using a Knowledge Base, remember customer information across sessions, calculate loyalty discounts using code, and browse live websites.

On paper, that sounds like six features.

In reality, it meant connecting AgentCore Runtime, AgentCore Gateway, Lambda, API Gateway, Knowledge Base, Memory, Code Interpreter, Browser Tool, IAM, deployment, and CloudWatch into one working system.

When I first saw all of that, it felt like a lot. I knew what I was supposed to build, but I didn't yet understand how all the pieces were supposed to fit together.

This time, I had enough time to work through the project properly instead of rushing through it. And that gave me room to notice something I probably wouldn't have noticed otherwise: I was making the project harder by trying to understand everything at once.

Once I stopped doing that and started asking, “What is the next thing I need to understand?”, the project became much easier to work through.

This isn't a step-by-step guide. There are already many great guides for AWS and AgentCore. This is my experience of building the project, what confused me, what went wrong along the way, how the architecture eventually started making sense, and what I learned from putting all those pieces together.

And somewhere along the way, the word “agent” stopped feeling like just another AI buzzword to me. I could actually see what it meant in a working system.

I Knew What I Had to Build. I Just Didn't Know How Big It Would Feel
So What Was This Agent Actually Supposed to Do?
The Architecture Finally Made Sense When I Stopped Thinking in AWS Names
I Drew the Architecture Because I Needed to See It
The Six Capabilities Were Really Six Different Problems
- 1. Order Tracking: Getting Real Data Back to the Customer
- 2. Refunds: When the Agent Needs to Take an Action
- 3. Knowledge Base: The Model Knowing Isn't the Same as the Application Knowing
- 4. Memory: When a New Conversation Isn't Really Starting From Zero
- 5. Code Interpreter: Let the Model Decide, Let Code Calculate
- 6. Browser Tool: When the Information Isn't Inside Your Application
Having the Right Code Doesn't Mean Having Access to It
Deployment Wasn't the Finish Line
Six Tests Were Easier Than One Giant “Does It Work?”
The Part That Became Clear Only After Everything Worked
There Was Another Layer I Hadn't Thought About Enough: Operating the System
I Also Had to Learn When the Project Was Actually Finished
What This Project Actually Taught Me
A Small Thank You
I Would Love to Hear From You
🤝 Let's Stay Connected

I Knew What I Had to Build. I Just Didn't Know How Big It Would Feel

I want to start here because the finished project looks much simpler than the experience of building it. Once everything was working, I could look at the architecture and understand what each part was doing.

But at the beginning, I saw a long list of unfamiliar names: Amazon Bedrock AgentCore, Strands SDK, Lambda, API Gateway, AgentCore Gateway, Knowledge Base, Memory, Code Interpreter, Browser Tool, IAM, and CloudWatch. And then there was deployment, testing, and permissions.

Individually, none of those things sounded impossible. Together, they felt like a lot.

I think that is one of the hardest parts of learning cloud and AI systems. You can understand each word separately and still have no idea how the words are supposed to connect.

I initially wanted to understand the entire architecture before I started building. That didn't work very well. The more I tried to understand everything at once, the more overwhelming it became.

So I changed the question I was asking myself.

Instead of:

“How am I supposed to understand this entire project?”

I started asking:

“What is the next thing I need to understand?”

That became a much more manageable way to work.

And eventually, the pieces started connecting.

So What Was This Agent Actually Supposed to Do?

The project was a fictional customer support system. A customer could ask something, and the AI agent needed to understand the request and use the appropriate capability.

The six capabilities were:

Order tracking
Refund processing
Product and policy questions
Customer memory across sessions
Loyalty discount calculation
Live website browsing

That meant this wasn't just a chatbot that generated a response. The agent could actually interact with other parts of a system, depending on what the customer needed.

A simplified view of the flow was:

Customer
   ↓
AI Agent
   ↓
What does the customer need?
   ↓
Use the appropriate capability
   ↓
Get the result
   ↓
AI Agent
   ↓
Final response
   ↓
Customer

At first, that flow looks straightforward. But every capability underneath it needed different infrastructure, and that is where the project became interesting.

The Architecture Finally Made Sense When I Stopped Thinking in AWS Names

The biggest change in my thinking came when I stopped looking at the project as a list of AWS services. Instead, I started looking at what job each piece had.

The AI model was responsible for understanding the customer's request and deciding what to do based on the instructions and available tools. For the project, I used Amazon Nova 2 Lite through Amazon Bedrock.

The Strands SDK gave me the framework for building the agent and connecting its behavior with the tools and capabilities it needed.

Then there was the infrastructure that helped the agent actually do things. AgentCore Runtime was where the deployed agent ran, while AgentCore Gateway provided the connection between the agent and backend tools. API Gateway exposed the order-related backend operations through APIs, and AWS Lambda performed those backend operations.

The other capabilities had their own pieces as well. The Knowledge Base provided product and policy information that the agent could retrieve, while AgentCore Memory allowed useful customer information to be retrieved across sessions. Code Interpreter handled calculations using actual code, and the Browser Tool allowed the agent to interact with live webpages. Finally, CloudWatch helped me monitor the deployed runtime.

Once I looked at those services by their jobs, rather than their names, the architecture became much less intimidating.

I wasn't trying to memorize AWS. I was trying to answer:

“What does my application need to do, and what piece is responsible for doing that?”

That question helped me much more.

I Drew the Architecture Because I Needed to See It

I eventually made an architecture diagram in Excalidraw. I wanted it to show the story of a request, rather than just put every AWS service on a page.

The idea was simple: the customer asks something, the agent figures out what the customer needs, the appropriate capability is used, and the result comes back to the agent. The agent then uses that result to respond to the customer. The supporting AWS services sit underneath the capabilities that need them.

This is the architecture I built for the project. Drawing it this way helped me understand the system as a flow instead of a collection of services.

Looking at the finished diagram now, it feels obvious to me. But I think that is because I built the pieces underneath it. The diagram became easier to understand after I understood what each box was solving.

The Six Capabilities Were Really Six Different Problems

Once the overall architecture made sense, I could look at each capability separately. That made the project feel much more manageable because I was no longer trying to understand six different things at the same time.

1. Order Tracking: Getting Real Data Back to the Customer

If a customer asked:

“Where is my order?”

the agent needed more than a generated sentence. It needed actual order information.

I created an order-tracking Lambda and exposed the required operations through API Gateway. AgentCore Gateway then provided the connection between the agent and those backend operations.

So the flow was roughly:

Customer
   ↓
Agent
   ↓
AgentCore Gateway
   ↓
API Gateway
   ↓
Lambda
   ↓
Order information
   ↓
Agent
   ↓
Customer

That was one of the first times the difference between a chatbot and an agent workflow became clearer to me. The model wasn't inventing the order status. It was asking another part of the system for the information.

2. Refunds: When the Agent Needs to Take an Action

Refund processing was similar, but the idea was slightly different. The agent wasn't just retrieving information. It needed to trigger an operation.

I connected the refund functionality through AgentCore Gateway to a Lambda function. This made something important click for me: a model saying that something happened and a system actually performing that action are two different things.

If an agent tells a customer:

“Your refund has been processed.”

there should be an actual operation behind that statement. That operation needs to succeed, the result needs to come back, and the agent should respond based on that result rather than simply assuming the action worked.

That is a very different way of thinking about AI applications compared with a simple question-and-answer chatbot.

3. Knowledge Base: The Model Knowing Isn't the Same as the Application Knowing

For product and policy questions, I used an Amazon Bedrock Knowledge Base. This was where I got a much more practical understanding of RAG.

Imagine a customer asks:

“What is the return policy?”

I don't want the model to generate something that sounds like a return policy. I want it to retrieve the information that belongs to the application.

The basic flow becomes:

Customer question
       ↓
Retrieve relevant information
       ↓
Give the information to the agent
       ↓
Generate the response

That made me realize something important: a model can know many things while still not know my application's information.

The Knowledge Base gives the agent a source of information it can retrieve from. That is much more useful than simply hoping the model happens to know the right answer.

4. Memory: When a New Conversation Isn't Really Starting From Zero

I also added AgentCore Memory. The goal was to allow the agent to remember useful customer information across sessions.

I tested this with a simple example. In one session, I told the agent a customer's name and response preference. Then I started a separate session and asked it to retrieve that information from the previous conversation.

It was able to do that.

This was interesting because the second session wasn't simply continuing the first conversation. The system had to retrieve the stored customer context.

That helped me understand the difference between:

“The model can see the previous messages.”

and:

“The application has a mechanism for retrieving relevant customer context.”

It also made me think about the responsibility that comes with memory. In a real customer support application, you would need to think carefully about what information is stored, how long it is retained, who can access it, and how customer information is protected.

Getting memory to work is only one part of the problem.

5. Code Interpreter: Let the Model Decide, Let Code Calculate

The loyalty discount was one of the more interesting pieces because I actually made a mistake here.

The agent needed to calculate a customer's discount using their loyalty tier, available points, and order total. My first calculation logic was incorrect because I was treating loyalty points and dollar values incorrectly.

The model wasn't the problem. My calculation logic was.

I fixed the calculation and redeployed the agent.

That taught me something I think is easy to forget when working with AI: an AI application still contains normal software engineering problems.

You can still have incorrect formulas, incorrect assumptions, edge cases, configuration problems, and bugs. The fact that a model is involved doesn't remove any of those.

It also helped me understand why Code Interpreter was useful. The model could understand that a calculation was needed, while code could perform the actual calculation. For precise arithmetic, that separation made much more sense to me than asking the model to do everything itself.

6. Browser Tool: When the Information Isn't Inside Your Application

The final capability was live web browsing. I used the AgentCore Browser Tool and tested it by asking the agent to visit the Udacity website and return the page title.

This was different from the Knowledge Base because the information wasn't sitting inside my application. The agent had to interact with a live webpage.

That helped me see another side of tools: a tool isn't necessarily just a function that returns data. It can give the agent a way to interact with something outside the model's immediate context.

Having the Right Code Doesn't Mean Having Access to It

This was probably one of the most practical lessons of the project. There were times when the architecture made sense and the code was there, but something still didn't work. The reason was permissions.

The AgentCore runtime needed the right permissions to access resources such as Memory and the Knowledge Base. The Browser Tool also needed the appropriate permissions before it could start a browser session.

This changed the way I approached debugging. Before this project, I might have immediately looked for a code problem. During this project, I started asking a different set of questions: Does the resource exist? Is it in the correct region? Does the runtime have permission to access it? Is the IAM role correct? Is the specific action allowed? Is the tool configured correctly?

That was a much better debugging mindset because sometimes the problem isn't:

“My code doesn't work.”

It is:

“My code is trying to use something it isn't allowed to use.”

The Browser Tool gave me a good example of this. I had the browser capability configured, but the runtime couldn't start the browser session. It would have been easy to assume that the Browser Tool itself was broken.

The actual issue was that the runtime was missing the required permission. I added the permission and tested it again, and this time, it worked.

It was a small fix, but it taught me something important about cloud applications: there is a lot around the code. The application has identities, permissions, resources, configurations, and service connections. All of those pieces have to agree before a feature can actually work.

Deployment Wasn't the Finish Line

Once the pieces were working, I deployed the agent using the AgentCore deployment workflow. This was another point where I changed how I think about deployment.

A successful deployment doesn't automatically mean a successful application. The runtime can deploy successfully while one of the tools still has a configuration or permission problem.

So I treated deployment as the beginning of testing rather than the end of development. I tested the deployed agent capability by capability, which gave me much more confidence than simply seeing a successful deployment message.

Six Tests Were Easier Than One Giant “Does It Work?”

Instead of asking myself, “Does my AI agent work?”, I broke that question into smaller ones.

Can it track an order? Can it process a refund? Can it answer from the Knowledge Base? Can it remember information across sessions? Can it calculate the loyalty discount correctly? Can it browse a live website?

Each one became its own test.

I captured evidence for each of these tests, which gave me six concrete things to verify instead of one enormous pass/fail question.

And eventually, all six worked.

That was a much less intimidating way to test the project.

The Part That Became Clear Only After Everything Worked

Looking back, I think the biggest shift happened after I had all six capabilities working. Before that, I was thinking about individual services. After that, I could finally see the pattern.

The model understands the request, the agent determines what it needs, and a tool provides the appropriate capability. Another service performs the work, the result comes back, and the agent uses that result to respond. That pattern repeats across the six capabilities.

The actual tools are different, but the underlying idea is similar:

Order tracking
→ retrieve backend information

Refund
→ perform a backend action

Knowledge Base
→ retrieve application information

Memory
→ retrieve customer context

Code Interpreter
→ perform computation

Browser
→ interact with live information

That was when the word agent finally started making sense to me.

It wasn't simply:

“An AI model that talks to you.”

It was:

“A model that can understand what needs to happen and use capabilities to actually do something about it.”

There Was Another Layer I Hadn't Thought About Enough: Operating the System

Once the agent was working, I started thinking about what would happen if this were a real customer support application. A customer doesn't care which service failed. They just know their request didn't work.

For example, an order request could travel through:

Customer
   ↓
AgentCore Runtime
   ↓
Agent
   ↓
AgentCore Gateway
   ↓
API Gateway
   ↓
Lambda
   ↓
Response

If something fails somewhere in that chain, the engineering team needs to know where. That's where monitoring becomes important.

For this project, I used CloudWatch to monitor the AgentCore runtime and created a CPU usage alarm. If I were taking this system further for real production usage, I would want to monitor things such as failed requests, error rates, response times, tool failures, resource usage, service health, and costs.

Cost would also matter because this agent uses multiple managed services. A project that works perfectly with a small number of requests could have a very different cost profile at much larger scale.

That made me realize something:

Building something that works and operating something reliably are different problems.

I Also Had to Learn When the Project Was Actually Finished

After the agent was working and the project was submitted, I went back through the AWS resources I had created and cleaned up the project-specific resources.

This wasn't the exciting part of the project, but I think it is an important part of cloud work. Getting the application to work isn't necessarily the same as being finished with the cloud environment around it.

When you create cloud resources, you should know what you created, what is still running, and what happens when you are done with it. Leaving resources active after a project is over can lead to unnecessary usage and costs, especially when multiple managed services are involved.

For someone learning with a limited cloud budget, that matters even more. I wanted to finish the project knowing that I had cleaned up what I no longer needed.

For me, cleanup became part of what actually finishing the project meant, not something separate from it.

What This Project Actually Taught Me

If I could go back to the beginning, I wouldn't tell myself, “Learn everything first.” I would tell myself:

“You don't need to understand the whole architecture before you start.”

When I first saw the project, I wanted to understand AgentCore, Lambda, API Gateway, IAM, RAG, Memory, Browser Tool, Code Interpreter, deployment, and monitoring all together. That was too much. I could understand what each service was called, but that didn't mean I understood how all of them were supposed to work together.

What helped was breaking the project down into smaller questions: What does this resource do? Why does the agent need it? What does it connect to? What should happen when I test it? What happens if it fails?

Those questions were much easier to answer, and each answer made the next one easier.

I think this is especially useful when you're starting something that looks too big. It is easy to look at a cloud architecture full of services and think:

“I don't know any of this.”

But you don't need to know all of it at once. Start with the application and ask what the user needs to happen. Then ask what the system needs in order to make that happen, and learn the service that solves that particular problem.

If you need to store data, learn the database piece. If you need to execute backend logic, learn the compute piece. If the agent needs to retrieve information, learn the retrieval piece. If it needs to remember customer context, learn the memory piece. If it needs to interact with an external capability, learn the tool or gateway piece.

You also don't have to assume that every problem is an AI problem. A missing permission can look like a broken feature. A configuration problem can look like broken code. A calculation bug can look like a model problem. A deployment can succeed while one capability is still misconfigured.

That was another important lesson for me. Before this project, I might have immediately looked for a code problem. During this project, I became more comfortable asking whether the resource existed, whether it was in the right region, whether the runtime had permission to access it, whether the IAM role was correct, and whether the specific action was allowed.

I started this project thinking I was mainly going to learn Amazon Bedrock AgentCore. I did, but I ended up learning several things around it too. I learned how an AI agent can use tools to interact with backend systems, how a Knowledge Base can give an agent access to application-specific information, how Memory changes what a conversation can mean across sessions, and why precise calculations sometimes belong in code rather than in the model's reasoning.

I also learned that IAM isn't something you set up once and forget about, that a successful deployment doesn't prove every capability works, that monitoring matters even for a relatively small project, and that cloud development includes knowing how to clean up what you create.

I don't think the biggest lesson I am taking into my next project is one specific AWS service. It is a way of approaching projects that initially look too big.

When I see a complicated architecture now, I don't immediately think:

“I need to understand everything.”

I think:

“What is the first piece?”

Then I work from there.

That doesn't mean the project suddenly becomes easy. There are still confusing moments, permissions to figure out, code to debug, and configuration to understand. Sometimes you understand a service after reading about it. Sometimes you understand it after using it. And sometimes you understand it only after something breaks and you have to figure out why.

All three are learning.

So if you're somewhere at the beginning of your own AI or cloud journey and you're looking at a project that feels like too much, start with one thing. Understand what it is supposed to do. Make that part work. Then move to the next.

You don't have to build the entire system in your head before you build it in reality.

And that is probably the thing I value most from this project.

I learned how to make a complicated project feel manageable.

Not by making the project smaller.

By making the next problem smaller.

A Small Thank You

I want to thank Udacity and everyone involved in the Future AWS Agent Engineer Nanodegree Program for giving me the opportunity to work through a project like this. And I'm especially grateful for the AWS AI & ML Scholarship that made the program possible for me.

This project gave me something I don't think I would have understood as clearly by only reading about AgentCore. I had to actually connect the pieces, troubleshoot them, and test them. Eventually, I could look back at the architecture and realize that I understood what I was looking at.

That was probably the most valuable part for me.

I Would Love to Hear From You

If you've built an AI agent, worked with AWS, or are currently learning cloud engineering, I'd genuinely love to hear from you.

What was the first cloud or AI concept that felt completely confusing at first, but eventually started making sense to you? Or, if you're currently building something difficult, what part are you working through right now?

Share it in the comments.

Sometimes the thing that feels obvious to you now is exactly the thing someone else is struggling to understand.

🤝 Let's Stay Connected

Place	Find me here
GitHub	building things → hemapriya-kanagala
LinkedIn	resources & updates → hemapriya-kanagala
X	random dev thoughts → @KanagalaHema

Dev Opportunity Radar #18: $50K AI Hackathon, SheBuilds, Verci Fellowship, and Stanford CS336

Hemapriya Kanagala — Fri, 25 Sep 2026 17:48:13 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the Verci Fellowship, Nebius x NVIDIA Global AI Hackathon, and SheBuilds on Lovable, along with Stanford CS336: Language Modeling from Scratch, a course for understanding how language models are built from the ground up.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

This week's 🌟 Community Finds section features two opportunities shared by Dhruv Jani (@dj29). Thank you, Dhruv, for sharing the Snapdragon® AI Lab Build & Present Challenge and the SerpApi India Hackathon 2026. I always enjoy seeing readers help others discover opportunities they might otherwise have missed, and I hope this section continues to grow.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, attended an event, joined a community, completed a program, built something, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, event, or anything else you think deserves more attention, feel free to share it in the comments.

If I feature one of your 🌟 Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Verci Fellowship
- 📌 Nebius x NVIDIA Global AI Hackathon
- 📌 SheBuilds on Lovable
📚 Resources Worth Checking Out
- Stanford CS336: Language Modeling from Scratch
🌟 Community Finds
- 📌 Snapdragon AI Lab Build & Present Challenge
- 📌 SerpApi India Hackathon 2026
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Format	Deadline
Verci Fellowship	Verci	Fellowship	Hybrid	October 1
Nebius x NVIDIA Global AI Hackathon	Nebius × NVIDIA	Hackathon	Online	October 30
SheBuilds on Lovable	Lovable	Build Event	Virtual	September 27

Resource Highlight

Resource	Why Check It Out
Stanford CS336: Language Modeling from Scratch	An advanced course covering the process of building language models from the ground up, including Transformers, training, GPU optimization, scaling, data, evaluation, and alignment.

Community Finds

Shared By	Find	Type	Format
Dhruv Jani (@dj29)	Snapdragon AI Lab Build & Present Challenge	AI Challenge	Online
Dhruv Jani (@dj29)	SerpApi India Hackathon 2026	Hackathon	Online & Async

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications.

So I keep this section to the two most recent editions as a quick reminder of opportunities you may have missed. Older opportunities are still available on the Dev Opportunity Radar website, where you can browse the full collection and use the filters to find something specific.

Opportunity	Organization	Type	Format	Deadline	Featured In
Code in Place X	Stanford University	Programming Course	Virtual	September 28	Edition #17
Build, Ship, Shape: Amazon Developer Hackathon	Amazon	Hackathon	Online	October 24	Edition #17
Dev3Pack Hackathon	Dev3Pack	Hackathon	Hybrid	October 30	Edition #17
AI for Education Fellowship	Worldwide Studios	Fellowship	Virtual (Global)	Rolling	Edition #16
AWS Student Rewards	AWS Builder Center	Student Rewards Program	Online (Global)	Open	Edition #16

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Verci Fellowship

Who it's for: Young builders, students, recent graduates, and people building projects or communities.

What it's about: The Verci Fellowship is a one-semester program bringing together 20 people who are building startups, projects, and communities.

The fellowship focuses heavily on community and mentorship. Each fellow is matched with a mentor, joins a small weekly mastermind group, and gets access to office hours and workshops with founders, operators, and investors.

There is also a Catskills retreat in November where the cohort spends a weekend together, along with support for fellows who want to organize things in their own communities such as hackathons, dinners, or demo nights.

Another part I found interesting is the focus on building in public. The program includes workshops and accountability groups to help fellows share what they are working on and actually ship something during the semester.

The fellowship is open to people who are currently in college, on leave, dropped out, or recently graduated, with the program looking for people roughly under 24 who have a strong interest in technology, are creative and curious, and are already building things on campus or in their communities.

Applications are rolling, but the current application form says applications close on October 1, 2026 at midnight PST.

Format: Hybrid / Online + New York and LA opportunities

Cohort: 20 fellows

Program: October – December 2026

Application Deadline: October 1, 2026 at midnight PST

Eligibility: Currently in college, on leave, dropped out, or recently graduated; roughly under 24

What you get: Individual mentorship, weekly mastermind groups, workshops and office hours, Catskills retreat, community/event support, cohort Slack, building-in-public support, and optional opportunity matching

Application: Name and background + social profiles + semester schedule + work/portfolio + short answers + 1–2 minute personal video

🔗 Apply to the Verci Fellowship

📌 Nebius x NVIDIA Global AI Hackathon

Who it's for: AI builders, ML engineers, developers, founders, and anyone interested in building practical AI systems.

What it's about: The Nebius x NVIDIA Global AI Hackathon is a global online hackathon where participants build an AI project using NVIDIA open source models on Nebius Token Factory or Nebius AI Cloud.

There are four main tracks, so you can choose the type of project that fits what you want to build:

Coding and Agentic Engineering: Build coding agents and developer tools.
Best Apps and Agents: Build an application or agent designed to solve a real problem.
Personal AI: Build a private, always-on AI assistant with memory, skills, and access to your chosen tools and information.
Physical AI: Build robotics, IoT, or other systems that can sense and act in the physical world.

The main requirement is that your submission runs on Nebius Token Factory or Nebius AI Cloud and uses at least one NVIDIA open source model.

I wanted to include this one because there is quite a bit of room in what you can build. You can work on an AI agent, developer tool, personal assistant, application, robotics project, or something else that fits one of the tracks.

There are also Builders & Brews events in different cities where participants can join technical onboarding, workshops, team formation, build sessions, and get feedback while working on their projects.

For the submission, you'll need a working project, public code repository, project description, working demo or test build, and a public demo video of three minutes or less. Your repository also needs an open source license and setup instructions.

There is $50,000+ in prizes, including a $20,000 grand prize, $10,000 for second place, $6,000 for third place, track prizes, city awards, and additional prizes.

Format: Online | Global

Deadline: October 30, 2026 at 10:30 PM IST

Tracks: Coding & Agentic Engineering | Best Apps & Agents | Personal AI | Physical AI

Requirement: Use Nebius Token Factory or Nebius AI Cloud + at least one NVIDIA open source model

Submission: Working project + public repository + demo + ≤3-minute demo video

Prize Pool: $50,000+ in prizes

🔗 Join the Hackathon

📌 SheBuilds on Lovable

Who it's for: Women 18+ who have an idea they want to turn into a working product, including students, first-time makers, founders, and experienced developers.

What it's about: SheBuilds on Lovable is a global community of women building with AI through virtual build events, webinars, and community spaces.

Season 04 is a 72-hour virtual build event where participants use Lovable to turn an idea into a working product alongside mentors, community champions, and a few hundred other builders.

You don't need to know how to code to apply. You can come with an idea, sketches, research, or a plan, and the actual building starts when the event kicks off.

You can participate solo or with one co-builder. Around 400 participants are expected to be selected for this season.

There are also live prep sessions from October 6–14 covering things like figuring out what to build, making your product stand out, and getting it in front of people. Accepted participants also get Lovable credits to use during the build.

The build starts on October 15 and submissions close on October 17. The organizers will announce the prizes at kickoff, so the prize details for Season 04 have not been announced yet.

Format: Online | 72-hour build event

Build Dates: October 15–17, 2026

Applications Close: September 27, 2026 at 11:59 PM PT

Decisions: October 5, 2026

Eligibility: Women 18+ with an idea they want to build

Team Size: Solo or 1 co-builder

Cost: Free | Coding Required: No

Expected Cohort: Around 400 participants

Prep Sessions: October 6–14

Prizes: To be announced at kickoff

🔗 Apply for Season 04

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

Stanford CS336: Language Modeling from Scratch

If you want to go beyond using LLMs and actually understand how language models are built, Stanford's CS336: Language Modeling from Scratch is a really interesting resource to explore.

The course walks through the process of building language models from the ground up, including data collection and cleaning, tokenization, Transformer architecture, training, GPU optimization, scaling, evaluation, inference, alignment, and reasoning.

What makes this course different is how implementation-heavy it is. The assignments have students actually build components of a Transformer language model, train a small language model, optimize attention with Triton, work with distributed training, process pretraining data, and apply post-training methods such as supervised fine-tuning and reinforcement learning.

The course provides lecture recordings, assignments, lecture materials, and code, so you can follow along even if you're not enrolled at Stanford.

One important thing to know before jumping in: this is an advanced course. The course expects proficiency in Python, experience with deep learning and systems optimization, familiarity with PyTorch, and a foundation in calculus, linear algebra, probability, statistics, and machine learning.

You will also be writing a lot of code, so this is much more suitable if you already have a solid technical foundation and want to understand the engineering behind language models.

Format: Self-paced online resource

Focus: Language models, Transformers, training, systems optimization, scaling, data, evaluation, alignment, and reasoning

Level: Advanced

Prerequisites: Python + PyTorch/deep learning + systems optimization + calculus + linear algebra + probability/statistics + machine learning

Includes: Lectures + recordings + assignments + code + course materials

🔗 Explore the CS336 Course Website

🎥 Watch the Lecture Recordings

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

This week, Dhruv Jani (@dj29) sent me two opportunities that I thought were worth including:

📌 Snapdragon AI Lab Build & Present Challenge

Shared by Dhruv Jani

The Snapdragon® AI Lab Build & Present Challenge by Qualcomm is an individual online challenge for participants in India to build or propose an AI solution designed, developed, or optimized for Snapdragon-powered HP PCs.

You can submit a new project or significantly modify an existing project by adding AI models from Qualcomm AI Hub or other open-source platforms.

The challenge is open to Indian residents aged 18+ who own a Snapdragon-powered laptop.

Format: Online | Individual

Deadline: September 30, 2026 at 11:59 PM IST

Eligibility: India residents, 18+

Requirement: Snapdragon-powered laptop

Rewards: HP laptops, Qualcomm internship opportunity for the first award subject to eligibility, executive mentorship, and showcase opportunities

🔗 Learn More & Apply

📌 SerpApi India Hackathon 2026

Shared by Dhruv Jani

The SerpApi India Hackathon 2026 is an online and asynchronous hackathon where participants can build an app, AI agent, or open-source integration using live search data from SerpApi.

You can participate solo or in a team of up to five, use any programming language or framework, and work from anywhere in India. Existing projects are also allowed if the submitted version makes meaningful use of SerpApi.

There are six project tracks, and AI-assisted development is allowed.

The total prize value is ₹3 lakh+, including cash awards, SerpApi credits, track awards, community awards, and credits for every valid submission.

Format: Online & asynchronous

Deadline: October 10, 2026 at 11:59 PM IST

Eligibility: India residents, 18+

Team Size: 1–5 members

Requirement: Meaningful use of SerpApi search data

Prize Value: ₹3 lakh+

Submission: Public GitHub repository + project description + demo video under 3 minutes

🔗 Submit Your Project

Thank you to Dhruv Jani (@dj29) for thinking of the Radar and sharing this one with the community. It really means a lot 💙

I'd love for this section to keep growing.

If you've come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, feel free to share it in the comments.

If I feature it in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post