Kedar Hardikar

I I build AI systems that actually go into production — not just notebooks and demos. Over the past year I've shipped LLM-powered applications for enterprise clients including Dell and Saint-Gobain: RAG knowledge assistants, document intelligence pipelines (summarization, redaction, extraction), a speech-to-speech voice assistant using Whisper, and FastAPI backends that tie it all together. What I'm focused on now: joining a team where I can go deeper on LLM agents, evaluation frameworks, and production AI — ideally at a company building AI-first products rather than adding AI as an afterthought. Stack: Python · LangChain · LangGraph · FastAPI · ChromaDB · PostgreSQL · Kafka · OpenAI API · HuggingFace