RAG & AI Application Engineering
Builds RAG / Agent applications with Python, LangChain, and LangGraph, covering PDF / HTML parsing, text chunking, embeddings, Chroma, hybrid BM25 and vector retrieval, RRF, and citations. Uses LLMs for structured query planning and improves recall through source grouping, candidate quotas, and backfilling, with versioned Hit@K and MRR evaluation and Cross-Encoder reranking comparisons. Experienced with tool calling, structured output, context management, Redis caching, fallback handling, and safety controls before and after model calls.