Applied AI / backend engineer who builds LLM products end to end. Based in Delhi NCR, open to remote roles.
- Zavyo: sole engineer on a multi-tenant WhatsApp + AI SaaS, covering backend, frontend, the RAG layer, billing and deployment. The production code is private; the showcase repo has the architecture and sanitized excerpts.
- offline-rag-stack: an air-gapped, self-hosted RAG system with a tested no-egress guarantee. A script probes DNS, TCP and HTTP from every container and fails if anything gets out.
- LiteLLM PR #45358: fixes Vercel AI Gateway embedding models being registered as chat models, so the proxy health check resolved them as chat. This is an open PR, not merged.
- Quotation Generation Agent: a FastAPI + React agent that turns price-sheet PDFs and a customer's requirements into a structured quotation.
Python, FastAPI, LLM agents and RAG, MongoDB, Qdrant, Docker, React