Jayant Saxena

I build end-to-end MERN applications and now focus on Generative AI engineering, turning retrieval-augmented and agentic ideas into production-ready systems. I’m hands-on with the full workflow—from designing LLM pipelines to deploying them so they deliver real value.

Recently, I built an agentic RAG pipeline using LangGraph and LangChain with multi-step retrieval, hybrid search (BM25 + HuggingFace dense vectors), and Groq LLM, deployed on Streamlit Cloud. I upgraded retrieval from standard cosine similarity to hybrid search to improve precision on ambiguous and multi-intent queries.

I also published an open-source npm CLI tool using Groq LLM to automate repository documentation, reaching 200+ weekly downloads. I enjoy building tools that make teams faster and documentation more reliable without sacrificing quality.

Earlier at ZOOCO, Pune, I developed full-stack applications from scratch, owned end-to-end architecture and deployment, and shipped real-time in-app chat using Socket.io for concurrent production traffic. I led code review initiatives, enforced best practices, and also designed and integrated a Google OAuth consultation booking system synced directly to veterinary calendars.