HimanshuJ16

Location: Delhi NCR, India (UTC+5:30)

Remote: Yes

Willing to relocate: Yes, from mid-2027

Technologies: TypeScript, Next.js, React, Node.js, Python, PostgreSQL, Prisma, WebSockets/Socket.io, AWS, GCP, Docker, React Native

Résumé/CV: https://himanshujangir.com (full resume by email)

LinkedIn: https://linkedin.com/in/himanshujangir16

GitHub: https://github.com/HimanshuJ16

Email: himanshujangir16@gmail.com

Full-stack engineer focused on real-time systems, 3 years freelancing. Built a tracking platform for 500+ vehicles (100K+ managed users), a Python algo trading engine for Nifty options, and Frank, an open-source package that stops AI coding agents from flattering you and claiming unverified work (20+ agent hosts supported). Looking for a founding engineer or full-stack role.

Ha. To be clear it's not strategy or alpha - it's don't double-fill your orders, don't leak future data into the backtest. You can still go broke, just not because of the code.

On too many skills degrading performance - I've got about 500 in one domain and I don't think count is the problem. Overlap is. Two descriptions that both plausibly match the same situation and the agent picks wrong. That happens at 20 skills just as easily.

What helped was making every skill say where it stops, and hand each excluded case to whichever skill does cover it. Writing down the boundary, not just the trigger.

Bit surprised testing didn't come up. If a skill ships a reference implementation you can test that, and at least check the cross-references resolve. Doesn't prove the prose is right, but it catches the rot you describe in section 3.