At Accenture Solutions, I architected and deployed a production RAG system for Pfizer that retrieves information across 1,000+ technical documents. It improved engineering teams’ access to information.
I optimized retrieval with hybrid search, vector embeddings, metadata filtering, and reranking, improving retrieval accuracy by 20% and response relevance by 25%. I also reduced end-to-end response latency by 150 ms through asynchronous processing and backend optimization.
On my Multi-Agent Test Case Generation System project, I built a five-agent LangGraph pipeline with a self-correcting loop and human review. It generated traceable test cases with 95–100% requirement coverage on the stated evaluations.
I’ve also built an NL-to-SQL system and a production RAG document intelligence system. I use evaluation methods including RAGAS and DeepEval, and hold Anthropic Claude Developer – Foundations and Microsoft Azure AI Fundamentals certifications.
