Fix RAG Pipeline – HIPAA-Compliant Medical Document Q&A API (LLM/FastAPI)

We are looking for an experienced AI/Backend Engineer to troubleshoot and optimize a partially developed RAG (Retrieval-Augmented Generation) pipeline that processes HIPAA-regulated medical documents, including clinical notes, insurance claims, and prior authorization letters. The application is built using FastAPI, LangChain, OpenAI, and Pinecone, with integration into a Node.js backend. While the system is functional, it is currently producing inaccurate retrieval results and, in rare cases, exposing raw PHI in generated responses. Addressing these issues is critical to ensuring both accuracy and HIPAA compliance. Scope of Work - Review and debug the vector embedding and retrieval pipeline. - Optimize chunk sizing, overlap settings, and metadata filtering to improve retrieval accuracy. - Update LLM prompts and response handling to ensure PHI is properly redacted before responses are returned. - Verify that FastAPI endpoints return clean, structured JSON responses, including confidence scores where applicable. - Confirm that HIPAA audit logging is functioning correctly and capturing all required request activity. - Test and validate the overall system to ensure reliable, compliant, and accurate output. Requirements - Proven experience with RAG systems, LangChain, OpenAI APIs, and Pinecone. - Strong knowledge of FastAPI and backend API development. - Familiarity with HIPAA compliance, PHI handling, and healthcare-related applications. - Ability to quickly identify root causes and implement effective solutions. The codebase is clean, documented, and ready for review. We are seeking a professional who can efficiently diagnose the issues, implement fixes, and verify that the system meets both accuracy and compliance requirements.