At PNC Bank, I build production AI systems, including RAG pipelines with FAISS, Pinecone, and embedding models that reduced document processing time by 75%. I’ve also tuned retrieval and model routing to keep latency below two seconds while reducing token spend.
I contribute to end-to-end AI delivery, from model validation and testing to production handoff. I’ve deployed agentic systems using LangGraph, LangChain, and MCP connectors, and supported AWS workloads handling more than 1,500 requests per minute.
At Abbott, I built and evaluated RAG pipelines over unstructured documents, reducing manual review time by 70%. I also developed model training and validation workflows and deployed models supporting more than 1,200 requests per minute.
Earlier, at State street and Cisco, I developed production analytics and machine learning solutions using Python and SQL. My work included distributed data processing, model evaluation, and anomaly detection across operational data.
