At PreciQube, I developed RAG pipelines with LangChain and FAISS for an AI-powered optics design and simulation system. I also integrated Claude, Gemini, and Groq APIs and added a physics validation layer to catch model hallucinations before they reached users.
I built and deployed FastAPI services that turn natural-language requirements into structured optical component configurations for visualization and simulation in the Next.js and React optics canvas.
For Medi-Chat, I engineered an agentic diagnostic flow with specialist routing and implemented secure REST API routes, session-based authentication, and persistent conversation history.
I also built web and AI projects including TrailBlaze, an autonomous building management system, and a multilingual fixed-deposit assistant. My work spans AI and LLM engineering, backend APIs, and full-stack applications.
