Generative AI, RAG and LLM evaluation, taking models from prototype to production in healthcare.
Python and FastAPI services with React and TypeScript front ends, built and deployed end to end.
I build retrieval-augmented assistants and agentic workflows that cite their sources, route low-confidence cases to people, and run in production.
Four years across healthcare, manufacturing and identity security.
I'm an AI Engineer at CVS Health, where I build Retrieval-Augmented Generation assistants and LangGraph agentic workflows on Azure OpenAI and Amazon Bedrock.
Before that, I built predictive-maintenance models on live plant sensor data at Harley-Davidson and access-risk models for identity governance at Saviynt.
My focus areas are grounded retrieval, LLM evaluation and human escalation for low-confidence answers. I also build the React and TypeScript front ends for these tools.
Responses are generated from retrieved, approved sources and returned with citations.
Golden datasets, RAGAS and DeepEval run whenever retrieval or prompts change.
Low-confidence or sensitive cases route to a review queue.
Three roles, newest first. Hover for highlights; click for the full results.
Hover a project for a quick preview, click it for the full case study.
Languages, frameworks and platforms I work with.
PDF · experience, skills and certifications.
Open to conversations about GenAI, RAG, agentic systems and LLM evaluation. Email is the fastest way to reach me.