Deepan Chakravarthi

Experience

1. Lead Ai Engineer

   Tiger Analytics

   May 2024 — Present · Chennai, IN

   Leading AI engineering initiatives delivering production-grade Agentic AI and Generative AI systems across enterprise clients. Built production Text-to-SQL system using LangGraph agentic framework with Prompt Engineering and OpenAI GPT-4 integration — reduced false negatives by 20% through novel query-type-specific evaluation methodology Developed MCP (Model Context Protocol) server enabling AI agents to perform Entity Extraction, query external APIs, local databases and unstructured documents — deployed via FastAPI and REST APIs for seamless LLM workflow integration Implemented end-to-end RAG (Retrieval Augmented Generation) pipeline using FAISS, ChromaDB and Pinecone vector databases for high-accuracy enterprise knowledge retrieval Fine-tuned LLMs using LoRA/PEFT techniques for domain-specific accuracy improvements using HuggingFace model hub Built multi-modal attribute extraction system combining nearest-neighbour search (cosine similarity), HuggingFace models and Visual Question Answering — achieving 90% recall across large product catalogues Led cross-functional team of 9 (6 data scientists + 3 MLEs) using Agile/Scrum methodology to deliver end-to-end AI content generation pipeline with LLM-based content validation Implemented GitHub Actions CI/CD pipelines for automated model testing, validation and deployment across environmentsSkills: Generative AI · Agentic AI · LLMs · LangGraph · LangChain · OpenAI · Prompt Engineering · Entity Extraction · RAG · Vector Databases · FAISS · ChromaDB · Pinecone · MCP · HuggingFace · PyTorch · FastAPI · REST APIs · Python · MLOps · CI/CD · GitHub Actions · Agile / Scrum.