AI/ML Engineer | RAG & Agentic AI Systems
Building production-grade AI backends — RAG pipelines, agentic workflows, and LLM-powered APIs.
I'm an AI/ML Engineer building document-grounded RAG systems, agentic workflows, and production backends. Currently at Chinar Quantum AI, where I architect AI assistants used company-wide.
My work spans:
- RAG pipelines (vector search, reranking, streaming)
- Agentic AI development and rapid prototyping
- FastAPI backend architecture and API design
- LLM integration (Claude, Gemini, OpenAI)
- Vector databases and semantic search
- Auth, rate limiting, and production observability
- Building document-grounded RAG systems with reranking and citation
- Architecting agentic AI products end-to-end
- Designing secure, scalable FastAPI backends
- Improving retrieval precision and latency in production
RAG LangChain Agentic AI Claude API Gemini API OpenAI API Hugging Face
Pinecone ChromaDB FAISS Semantic Search Embeddings
Supabase JWT Auth WebSockets Stripe,ollama
FastAPI Machine Learning Deep Learning DockerKubernetes
Unified FastAPI backend combining a document-grounded RAG QnA system and a ticket-routing engine into one AI assistant, adopted company-wide at Chinar Quantum AI. Role-based access control, Pinecone vector search with cross-encoder reranking, and WebSocket streaming with Redis-backed memory.
Production-ready RAG system for financial document analysis — 89.3% retrieval precision@3, 94.1% answer accuracy, sub-500ms end-to-end query latency. 5-stage retrieval pipeline with cross-encoder reranking and cited response generation.
Backend for a SaaS platform that tailors resumes to job descriptions and runs scored AI mock interviews via Gemini. Three-tier freemium billing via Stripe, composable rate-limiting middleware, and Row-Level Security in Supabase Postgres.
Full-stack gear/trip-matching platform shipped by a 3-person team in a 1–4 week sprint — FastAPI, SQLAlchemy, React, Supabase/RLS, Redis, Docker, Claude API. Deployed on Render and Vercel.
What I value in every project:
- Clean, well-architected backend code
- Measurable retrieval and latency performance
- Defense-in-depth security (RBAC, RLS, rate limiting)
- Shipping fast without cutting corners
Open to remote, hybrid, and onsite roles. Always up for a conversation about System Design,RAG systems, agentic AI, or backend architecture and Engineering.