AI testing storyteller. Writing the 36 Stratagems series — ancient Chinese military tactics meet modern AI system meltdowns. 16yr QA → building AI test framework.
Hey, I'm Hema 👋
Developer, writer, and creator of Dev Opportunity Radar, a weekly series published every Friday on DEV, helping people discover opportunities they might otherwise miss.
Funny how things come full circle 🥹 It all started with one of these Weekly Win threads, and last week, Dev Opportunity Radar reached 25 editions!
I wanted to share it here too, now that I have finally finished putting together a special post looking back at the journey. If you would like to check it out, here it is:
I don't think I would have made it this far without the community's support. Thank you everyone. I'm really happy to share this little milestone with you all ❤️
Civic technologist building AI tools that help systems explain themselves. Clew Suite builder, NACM 2026 speaker. I write about hackathons, product judgment, agent readiness, and learning in public.
Hey, I'm Hema 👋
Developer, writer, and creator of Dev Opportunity Radar, a weekly series published every Friday on DEV, helping people discover opportunities they might otherwise miss.
Civic technologist building AI tools that help systems explain themselves. Clew Suite builder, NACM 2026 speaker. I write about hackathons, product judgment, agent readiness, and learning in public.
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
Hey, I'm Hema 👋
Developer, writer, and creator of Dev Opportunity Radar, a weekly series published every Friday on DEV, helping people discover opportunities they might otherwise miss.
Founder & CEO at AGI Systems Directorate | Full-Stack AI Builder
Architecting multi-hybrid, bio-inspired systems to solve real-world problems—building toward AGI and figuring out intelligence
This is the first time I am commenting on this post, though I have always seen it around! Seriously, I have multiple wins this week.
First of all, I finally sorted out the issues with my GitLab account for good.
Second, I am heading to the DeepMind headquarters in London for the second time for a meeting with the head of DeepMind. Seriously, that is a huge deal for me and I still cannot believe it is actually happening.
Third, and the biggest win of all, the full application for Project Lumina is finally ready! The build is done, only the release is left, and we are going to ship it very soon.
For a quick intro since I am commenting here for the first time: I am the Founder and CEO of AGI Systems Directorate. We mostly work to build AGI through a four-hybrid approach combining large language models, autonomous agents, robotics, and biological human brain replication mechanisms.
Civic technologist building AI tools that help systems explain themselves. Clew Suite builder, NACM 2026 speaker. I write about hackathons, product judgment, agent readiness, and learning in public.
Founder & CEO at AGI Systems Directorate | Full-Stack AI Builder
Architecting multi-hybrid, bio-inspired systems to solve real-world problems—building toward AGI and figuring out intelligence
Thanks! Honestly, I still cannot believe I am heading back to the DeepMind headquarters in London for a second time, it feels pretty surreal.
And yeah, when I first joined DEV, I seriously did not comment on any posts at all. I only really started commenting around September, and even then I had never jumped into this weekly thread before because I was a little bit afraid to, I do not even know why! Glad I finally did though.
Civic technologist building AI tools that help systems explain themselves. Clew Suite builder, NACM 2026 speaker. I write about hackathons, product judgment, agent readiness, and learning in public.
A second trip to DeepMind HQ is huge! I'd love to hear that story. And I'm really glad you jumped into this thread. Your comments are a big part of what makes a place like this feel like a community.
Founder & CEO at AGI Systems Directorate | Full-Stack AI Builder
Architecting multi-hybrid, bio-inspired systems to solve real-world problems—building toward AGI and figuring out intelligence
Actually, the story is pretty weird and honestly hard to believe, even though it really happened! It all started with a random reply on X (formerly Twitter) a few months ago. For the first time in my life, I commented on a post without even looking at the username or who was behind it, which turned out to be the head of DeepMind. Somehow, they actually replied back.
From there, the conversation moved to DMs, then voice calls, and eventually video calls. On one random video call, they suddenly asked, "So when are you coming to the London DeepMind headquarters to collaborate with me?" They were completely serious! So I flew out there on October 7th, and now I am heading back again on November 23rd.
Civic technologist building AI tools that help systems explain themselves. Clew Suite builder, NACM 2026 speaker. I write about hackathons, product judgment, agent readiness, and learning in public.
Wait, what?! A random reply on X turned into a trip to DeepMind HQ, and now a second one? That's a post waiting to happen. And it's such a good example of just showing up and joining the conversation without worrying about who's on the other end. Safe travels on the 23rd!
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
AI testing storyteller. Writing the 36 Stratagems series — ancient Chinese military tactics meet modern AI system meltdowns. 16yr QA → building AI test framework.
Civic technologist building AI tools that help systems explain themselves. Clew Suite builder, NACM 2026 speaker. I write about hackathons, product judgment, agent readiness, and learning in public.
AI testing storyteller. Writing the 36 Stratagems series — ancient Chinese military tactics meet modern AI system meltdowns. 16yr QA → building AI test framework.
AI testing storyteller. Writing the 36 Stratagems series — ancient Chinese military tactics meet modern AI system meltdowns. 16yr QA → building AI test framework.
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
Also finished writing 10 blogs in this entire week. Started writing regularly on DEV.to after nearly two years and trying to note down my thoughts and experiments going forward.
Experimented with different models like Claude opus 5.5 and fable 5.2 for threejs experiments, Google Gemini Hydrogen for a similar experiment, and Worldlabs Atlas tool.
Finished some of my side projects and fun hobby projects.
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
My win this week was turning some ideas that had been floating around for a while into something more concrete. It’s always satisfying when a project starts moving from “I think this could work” to actually taking shape.
Nothing dramatic, but progress is progress — and sometimes getting unstuck is the biggest win of the week. 🎉
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
Civic technologist building AI tools that help systems explain themselves. Clew Suite builder, NACM 2026 speaker. I write about hackathons, product judgment, agent readiness, and learning in public.
Backend developer turned builder. 20 years in software, now building Reporails - instruction governance for coding agents. Writing about context engineering and patterns that make AI agents reliable.
I got a negative feedback! Instead of defending myself or attacking the person, I turned it into an opprtunity, learned from it, made a new friend, and got better! :)
Full-stack dev shipping Django/React SaaS and Shopify stores since 2018. 37 client projects, 11 live Shopify builds. I write production war stories so you don’t repeat my outages.
Sometimes I code, always I write — A simple man making his way through the galaxy, writing as I go — 📚 Author of "Street-Smart Coding in the Age of AI": bit.ly/csarag-sscai
Location
Colombia 🇨🇴 (not Columbia)
Work
Content, Courses & Training for .NET teams — Helping teams to write maintainable & performant code
My win: I finally measured whether my coding agent needs a skill to find past decisions among ~500 ADRs. In 150 blind runs, correctness was already at the ceiling without it, and the skill only cut cost by 20%. Two thirds of that came from one sentence in CLAUDE.md telling the agent not to read the index files. The scripts mostly made that instruction stick: with the sentence alone, 8 of 50 runs still read the index; with the skill, 1 of 50.
I also pointed my reverse-architecture skill at karasu itself, and the architecture modeling tool drew its own architecture.
Small one this week. Two of the free tools on my site had been quietly broken: one threw a JavaScript error on every load, and the other never drew the dots on its two charts 😅
The first was an edit field bound to a habit that wasn't being edited. The second was an x-for inside an inline SVG. Both work now.
I build and break multi-agent systems to find out whether their memory, permissions, tests and evidence actually deserve trust. Everything I publish ships with receipts you can run yourself.
This week I published a long piece on the AI and data center debate ive been working on for a week. It starts with the movies we all grew up on, then follows that fear into what the CEOs, the bills, and the protests actually say, reading the whole thing instead of the half that usually gets quoted. We took our time with it and did the best we could to get the right information in.
Small one: my side project I Did My Own (I run it) collects open-source apps where someone rebuilt a famous tool, and this week I started writing up the projects on here. So far there's an offline take on Tinkercad and a roundup of four open-source stop-motion tools. Next goal is getting more makers to submit their own rebuilds.
¡Aloha! Gracias. Puedes ver mi último artículo en mi blog de DEV; ahí están los últimos avances e hitos. :)
Aloha! Thanks. You can check out my latest article on my DEV blog, where you'll find the latest updates and milestones. :)
Aloha ! Merci. Tu peux consulter mon dernier article sur mon blog DEV, tu y trouveras les dernières avancées et jalons. :)
Y Jess, con respecto al anuncio que pusiste hace poco sobre la oferta de trabajo en DEV yo te recomiendo contratar a Hema, habla con ella, veras que le va a encantar la idea, ella siempre esta ayudando mucho a todos en la comunidad DEV y es muy activa y muy buena onda, aqui te dejo su blog: dev.to/hemapriya_kanagala
"And Jess, regarding the job opening you posted recently on DEV, I highly recommend hiring Hema. Reach out to her, you'll see she's going to love the idea! She's always helping everyone out in the DEV community, she's super active and has a great vibe. Here's her blog: dev.to/hemapriya_kanagala"
Not a Win from me, but a Win for @xulingfeng for getting this badge!
When will the Devengers web series come out?
oh sorry i mean website
We have a Website, but it is currently counting down the-devengers-tower.github.io/
55days!
Hahaha🤣, I’m riding on the glow of your post! Thank you so much😆
Funny how things come full circle 🥹 It all started with one of these Weekly Win threads, and last week, Dev Opportunity Radar reached 25 editions!
I wanted to share it here too, now that I have finally finished putting together a special post looking back at the journey. If you would like to check it out, here it is:
Behind-the-scenes look and community AMA
25 Editions of Dev Opportunity Radar: From My Radar to Our Radar 💙 + AMA
I don't think I would have made it this far without the community's support. Thank you everyone. I'm really happy to share this little milestone with you all ❤️
Woohoo!
lol
Life is full of those full circle its amazing when we can appreicate them.
So true 😀
This is the first time I am commenting on this post, though I have always seen it around! Seriously, I have multiple wins this week.
First of all, I finally sorted out the issues with my GitLab account for good.
Second, I am heading to the DeepMind headquarters in London for the second time for a meeting with the head of DeepMind. Seriously, that is a huge deal for me and I still cannot believe it is actually happening.
Third, and the biggest win of all, the full application for Project Lumina is finally ready! The build is done, only the release is left, and we are going to ship it very soon.
For a quick intro since I am commenting here for the first time: I am the Founder and CEO of AGI Systems Directorate. We mostly work to build AGI through a four-hybrid approach combining large language models, autonomous agents, robotics, and biological human brain replication mechanisms.
Woah, what a week. Congrats!
Thanks! Honestly, I still cannot believe I am heading back to the DeepMind headquarters in London for a second time, it feels pretty surreal.
And yeah, when I first joined DEV, I seriously did not comment on any posts at all. I only really started commenting around September, and even then I had never jumped into this weekly thread before because I was a little bit afraid to, I do not even know why! Glad I finally did though.
A second trip to DeepMind HQ is huge! I'd love to hear that story. And I'm really glad you jumped into this thread. Your comments are a big part of what makes a place like this feel like a community.
Actually, the story is pretty weird and honestly hard to believe, even though it really happened! It all started with a random reply on X (formerly Twitter) a few months ago. For the first time in my life, I commented on a post without even looking at the username or who was behind it, which turned out to be the head of DeepMind. Somehow, they actually replied back.
From there, the conversation moved to DMs, then voice calls, and eventually video calls. On one random video call, they suddenly asked, "So when are you coming to the London DeepMind headquarters to collaborate with me?" They were completely serious! So I flew out there on October 7th, and now I am heading back again on November 23rd.
Wait, what?! A random reply on X turned into a trip to DeepMind HQ, and now a second one? That's a post waiting to happen. And it's such a good example of just showing up and joining the conversation without worrying about who's on the other end. Safe travels on the 23rd!
Create submit for VR competition under 11 days
( including a last 3 days Meta VR device )
100 x |> exhaused
very cool !
2 wins this week!
One Year on DEV 🎂
Heartfelt thank-yous and community shoutouts
One Year on DEV: The People Made the Platform
Congrats 🎉
✨ 🎉
first of many more to come, glad youre here!
very very congrats😍
Thanks ✨
The Thirty‑Six Stratagems series is officially complete!🎉
Nice! Congratulations!
😝Thankyou!
Nice, congrats! 🎉 When will you post the weekend update?
Hahaha! What weekend update? I know nothing about it 😂
I mean!🤔
After a long time, I participated in a DEV Community challenge and submitted a project I am satisfied with.
Presage SmartSpectra SDK Challenge Submission
Hearth: a desk companion that looks after you kindly
Thats a really neat video
Thank you
github.com/users/amananandrai/achi...
Received 128 github stars on a repo and Bronze level starstruck badge.
Also finished writing 10 blogs in this entire week. Started writing regularly on DEV.to after nearly two years and trying to note down my thoughts and experiments going forward.
Experimented with different models like Claude opus 5.5 and fable 5.2 for threejs experiments, Google Gemini Hydrogen for a similar experiment, and Worldlabs Atlas tool.
Finished some of my side projects and fun hobby projects.
getting 128 stars on github is very neat love to see it 🔥
Thanks
Every day and every week on dev.to is special ❤️I learn so many new things by reading articles from subject experts. It’s a great place to learn.
Yes that is very true getting to learn all sorts of new things on DEV is special indeed this community is amazing.
My win this week was turning some ideas that had been floating around for a while into something more concrete. It’s always satisfying when a project starts moving from “I think this could work” to actually taking shape.
Nothing dramatic, but progress is progress — and sometimes getting unstuck is the biggest win of the week. 🎉
love to hear it, keep brainstorming!
receiving beautiful badge!!
great feeling when you wake up to getting those badges its a very nice feeling. keep doing it big !
My win this week: I got the JavaScript badge for "One Month After the Hackathon"! The top #javascript post of the week. Absolutely crazy.
Improving the Reporails Xray functionality for context debugging and finalizing a robust Software Factory
Fun times ^^
I got a negative feedback! Instead of defending myself or attacking the person, I turned it into an opprtunity, learned from it, made a new friend, and got better! :)
I won a best use of browserbase track in hacktoberfest challenge on dev!!!!
Finally releasing the davaviz about AI racist bias :
vicentereyes.org
I just finished a podcast interview in the Write While True podcast about my writing process and my latest book.
I started a side quest : exploring stock markets and crypto
(and yeah, laundry too)
My win: I finally measured whether my coding agent needs a skill to find past decisions among ~500 ADRs. In 150 blind runs, correctness was already at the ceiling without it, and the skill only cut cost by 20%. Two thirds of that came from one sentence in CLAUDE.md telling the agent not to read the index files. The scripts mostly made that instruction stick: with the sentence alone, 8 of 50 runs still read the index; with the skill, 1 of 50.
I also pointed my reverse-architecture skill at karasu itself, and the architecture modeling tool drew its own architecture.
Write-up: dev.to/kompiro/one-sentence-in-cla...
This week, I was looking for a way to create polished demo GIFs without screen recording, retakes, or opening a video editor.
I just asked Claude to make one for DPOLens, and it:
The demo shows Claude Code checking a privacy policy through DPOLens before writing logging code, then citing the relevant clause in the code.
Everything on screen actually happened.
The GIF is in the repo's README if you want to see it: github.com/alkhatibdev/dpolens
Also, a small milestone that made my day: DPOLens received its first 5 stars yesterday ⭐
Small one this week. Two of the free tools on my site had been quietly broken: one threw a JavaScript error on every load, and the other never drew the dots on its two charts 😅
The first was an edit field bound to a habit that wasn't being edited. The second was an x-for inside an inline SVG. Both work now.
This week I published a long piece on the AI and data center debate ive been working on for a week. It starts with the movies we all grew up on, then follows that fear into what the CEOs, the bills, and the protests actually say, reading the whole thing instead of the half that usually gets quoted. We took our time with it and did the best we could to get the right information in.
Everybody Has an AI Apocalypse Story. Almost Nobody Reads the Fine Print.
Next up is the companion I'm building, so he can notice things on his own and come to me when something matters.
Small one: my side project I Did My Own (I run it) collects open-source apps where someone rebuilt a famous tool, and this week I started writing up the projects on here. So far there's an offline take on Tinkercad and a roundup of four open-source stop-motion tools. Next goal is getting more makers to submit their own rebuilds.
¡Aloha! Gracias. Puedes ver mi último artículo en mi blog de DEV; ahí están los últimos avances e hitos. :)
Aloha! Thanks. You can check out my latest article on my DEV blog, where you'll find the latest updates and milestones. :)
Aloha ! Merci. Tu peux consulter mon dernier article sur mon blog DEV, tu y trouveras les dernières avancées et jalons. :)
Y Jess, con respecto al anuncio que pusiste hace poco sobre la oferta de trabajo en DEV yo te recomiendo contratar a Hema, habla con ella, veras que le va a encantar la idea, ella siempre esta ayudando mucho a todos en la comunidad DEV y es muy activa y muy buena onda, aqui te dejo su blog: dev.to/hemapriya_kanagala
"And Jess, regarding the job opening you posted recently on DEV, I highly recommend hiring Hema. Reach out to her, you'll see she's going to love the idea! She's always helping everyone out in the DEV community, she's super active and has a great vibe. Here's her blog: dev.to/hemapriya_kanagala"
Starting a framework for a game I’ve always wanted to write.
I’ll post about it here in the coming days.