Description
Summary We are looking for a senior AI engineer to improve and expand our production Healthcare AI platform powered by Retrieval-Augmented Generation (RAG). The platform helps healthcare professionals and support teams quickly access trusted information from clinical documentation, medical guidelines, internal knowledge bases, and operational documents. Our current system works, but retrieval accuracy, response quality, and scalability need improvement. We're looking for someone who can take ownership of the RAG pipeline, enhance its performance, and build reliable AI automations that reduce manual work while ensuring responses remain grounded in source documents. This is a hands-on engineering role with architectural input—not a consulting-only position. What You'll Do Audit and improve the existing RAG architecture end-to-end Optimize document ingestion, parsing, chunking, and indexing for healthcare content Improve embedding quality, hybrid retrieval, metadata filtering, and re-ranking Reduce hallucinations through prompt engineering, context optimization, and LLM tuning Design AI automation workflows for document processing, knowledge management, and support operations Optimize vector database performance and retrieval latency Build evaluation pipelines to measure retrieval accuracy, answer quality, groundedness, and hallucination rates Implement source citations and confidence scoring Document the architecture and establish best practices for future development Required Skills Proven experience building production-grade RAG applications Strong Python development skills Experience with OpenAI, Anthropic Claude, or similar LLM APIs Deep understanding of embeddings, vector databases, semantic search, and retrieval optimization Experience implementing hybrid search (vector + keyword) and re-ranking models Hands-on experience with LangChain, LangGraph, LlamaIndex, or similar frameworks Strong prompt engineering and NLP fundamentals Experience building AI automation workflows Excellent written English and async communication Nice to Have Experience with healthcare, medical, or HIPAA-compliant AI applications Familiarity with FHIR, HL7, or healthcare data standards Experience with Pinecone, Weaviate, Qdrant, pgvector, or Milvus Knowledge of FastAPI, Docker, Kubernetes, and cloud platforms (AWS, Azure, or GCP) Experience building evaluation frameworks using Ragas, DeepEval, or similar tools Success Looks Like Higher retrieval precision and recall More accurate, source-grounded responses Significantly fewer hallucinations Faster retrieval and response times Reliable AI automation workflows that reduce manual effort A scalable, well-documented RAG architecture that can continue to evolve as our healthcare platform grows.