At Deloitte, I developed a healthcare Generative AI solution with AWS Bedrock, Claude, and LangChain that reduced manual document review by 45%. I also improved clinical and claims document retrieval accuracy by 32% through RAG search refinement.
I automated multi-step healthcare workflows with LangGraph, Semantic Kernel, ReAct, and MCP, and built integrations for HL7, FHIR, and EDI X12 data. These systems support document processing and access to clinical and claims information.
At UHG, I developed a clinical NLP and document processing platform using PyTorch and Hugging Face Transformers, reducing manual processing time by 30%. I also built claims classification and risk-prioritization models to help teams identify cases for earlier review.
Earlier, I developed real-time fraud detection for American Express, personalization and demand forecasting for Hilton, credit risk analytics for CSI, and retail data pipelines at EXL Service. My work spans production AI/ML, data engineering, and model deployment across healthcare, financial services, hospitality, and retail.
