● Built a multi-agent patient-support system using LangGraph, LangChain, and FastAPI to orchestrate tool-calling across inquiry resolution and healthcare knowledge retrieval workflows, reducing request handling from 6 steps to 2 automated actions. ● Designed a hybrid RAG pipeline using BM25, Azure OpenAI text-embedding-3-large, and Reciprocal Rank Fusion over 85K+ provider guidelines and ICD-10 documents, improving retrieval precision by 18 points over vector-only search. ● Built document-ingestion pipelines using Azure Document Intelligence for OCR preprocessing and metadata validation across 40K+ protected healthcare records from 3 enterprise repositories, reducing ingestion error rate from 12% to under 2%. ● Developed an LLM evaluation framework using Ragas, LLM-as-a-Judge, and MLflow across 120+ healthcare test scenarios, enforcing answer faithfulness scores above 0.85 as a hard production release gate. ● Configured CI/CD and monitoring pipelines using Azure Machine Learning and Azure AI Search across 3 healthcare business units, enabling model releases with automated drift detection and 4-hour rollback SLA.…● Built a multi-agent patient-support system using LangGraph, LangChain, and FastAPI to orchestrate tool-calling across inquiry resolution and healthcare knowledge retrieval workflows, reducing request handling from 6 steps to 2 automated actions. ● Designed a hybrid RAG pipeline using BM25, Azure OpenAI text-embedding-3-large, and Reciprocal Rank Fusion over 85K+ provider guidelines and ICD-10 documents, improving retrieval precision by 18 points over vector-only search. ● Built document-ingestion pipelines using Azure Document Intelligence for OCR preprocessing and metadata validation across 40K+ protected healthcare records from 3 enterprise repositories, reducing ingestion error rate from 12% to under 2%. ● Developed an LLM evaluation framework using Ragas, LLM-as-a-Judge, and MLflow across 120+ healthcare test scenarios, enforcing answer faithfulness scores above 0.85 as a hard production release gate. ● Configured CI/CD and monitoring pipelines using Azure Machine Learning and Azure AI Search across 3 healthcare business units, enabling model releases with automated drift detection and 4-hour rollback SLA.WWWW…