I build RAG systems, LLM applications, and multi-agent AI systems — from retrieval pipelines to production deployments.
I'm an MCA student at VIT Vellore focused on practical, production-grade AI. I design and ship GenAI systems end to end — hybrid retrieval, reranking, evaluation loops, APIs, caching, auth and deployment — rather than notebook demos.
A production-grade retrieval-augmented generation system combining BM25 sparse search and vector-based dense search (ChromaDB / AstraDB), merged with Reciprocal Rank Fusion and cross-encoder reranking (ms-marco-MiniLM-L-6-v2) for improved relevance. FastAPI backend, Redis caching with SHA-256 query hashing, Streamlit frontend with separate admin and public views, Firebase Google Auth, deployed via Docker on AWS EC2.
An AI-powered tool that processes meeting transcripts and recordings to generate concise summaries, key discussion points, and action items using an LLM pipeline — reducing manual note-taking and making meeting outcomes easy to review and share.
Vellore Institute of Technology, Vellore
University of Calicut
Open to AI/GenAI internships, research collaborations, and freelance builds. Drop a line — I usually reply within a day.