I've built an Agentic RAG document Q&A system that queries 103 PDF documents, including scanned files, and achieved a 94–95% observed successful answer rate in testing. Using Python, Gemini API, Qdrant, and Sentence Transformers, I designed a six-stage ingestion pipeline for source- and page-referenced responses.
I've also developed NLP and stock-analysis projects while bringing client-facing, CRM, and data-analysis experience from Aryodaya Hyundai, HDB Financial Services, and Raj Automotive Honda. I combine customer needs with practical Python, SQL, Core Java, machine learning, and generative AI work.

