At BestBuy, I build enterprise GenAI knowledge assistants using LangChain, LangGraph, RAG, and multi-agent workflows. I optimized semantic retrieval and re-ranking to reduce token consumption by 9.4%, and helped reduce operational overhead by over 20%.
I also built a Text-to-SQL agent and PromptHub, an internal LLM playground for experimenting with models, vector stores, tools, and agent configurations. I deploy containerized AI services on AWS using EC2, S3, Docker, and AWS Titan Embeddings.
Previously at Mastercard, I developed ML systems for AML, KYC, trade monitoring, and employee conduct oversight. My work reduced AML false positives by around 80% and manual investigation effort by 50%, while combining computer vision, NLP, RAG, and scalable data processing.
