I build with LLMs, RAG systems, and AI agents. Not the kind that demos well and breaks in prod. The kind that actually ships.
Trained neural networks. Experimented with transformers. Shipped RAG applications and agentic workflows that solve real problems for real people.
Right now I'm deep into evals and observability β figuring out how to measure if AI systems actually work and catching issues before they hit production. Always experimenting with new frameworks, always staying updated with what's actually worth using.
|
AI copilot for Indian law Semantic search and Q&A over 48K+ legal documents. Built for lawyers who need answers they can trust, not hallucinated citations. |
Automated GitHub issue management Reads code. Understands context. Suggests fixes. Because triaging issues at 2 AM shouldn't be a human's job. |
"The best code is the code you never had to debug at 3 AM.
The second best is the code that debugs itself."
β me, probably



