AI/ML Engineer at Databricks

Built 6+ production-grade Generative AI applications using Python, LangChain, LangGraph, and Databricks, serving 1,500+ enterprise users and processing 20,000+ monthly LLM requests. Designed scalable RAG pipelines across 12+ structured and unstructured enterprise data sources, improving response relevance by 28% and reducing retrieval time by 40%. Delivered AI agent and multi-agent workflows with planning, memory, tool calling, and human-in-the-loop validation, automating 8 business processes and saving 120+ hours per month. Improved semantic retrieval accuracy from 72% to 89% using Databricks Vector Search, hybrid search, metadata filtering, query rewriting, contextual retrieval, and reranking. Fine-tuned open-source LLMs using LoRA/QLoRA/PEFT, improving domain-specific accuracy by 18% while reducing GPU memory usage by 45%. Implemented automated LLM evaluation covering 500+ test scenarios and deployed secure AI controls (prompt injection protection, PII masking, RBAC, content filteri