I'm developing an AI-powered pharmaceutical decision intelligence system that analyzes drug candidates and generates structured Go, Conditional Go, or No-Go assessments. I built LLM workflows with Python, FastAPI, and Claude, and designed the production architecture around PostgreSQL, RAG, and vector search.
On a RAG-based document intelligence project, I developed a question-answering system for PDFs with document ingestion, semantic search, and context-aware responses using LangChain, ChromaDB, and Groq-hosted LLaMA models.
I also developed an AI recruitment screening and voice interview agent. Its resume screening combines rule-based skill matching, semantic similarity, and LLM evaluation, while its voice interview pipeline uses LiveKit, WebRTC, Whisper, Ollama, and text-to-speech.
