Building Production RAG Systems for Health Care Domains
summary
Claims concrete: a RAG corpus of more than 7,800 clinical publications and cross‑specialty validation, vector retrieval optimizations that deliver up to 35% latency gains, and a three‑tier evaluation that boosts critical‑error detection 2.4× are reported. Implementation uses Airflow nightly ingest, Selenium/BeautifulSoup scraping, Snowflake → Pinecone embeddings, LangChain hybrid retrieval, OpenAI and Gemini models, AWS Comprehend deidentification, and NeMo/guardrails; demo shows a working on‑cloud clinical RAG with nightly refreshes and claimed 60% inference cost reduction.
AI-written summary · about
chapters (16)
- 0:00AI challenges in healthcare
- 3:21Production RAG systems matter
- 8:39Healthcare AI regulations
- 10:05Healthcare RAG requirements
- 13:12RAG implementation challenges
- 16:02Healthcare RAG architecture
- 20:57Data pipeline and updates
- 22:57Hybrid search and chunking
- 25:04Tokenization and hallucinations
- 26:49Benchmarking healthcare RAG
- 27:33Security and compliance
- 29:25Healthcare RAG best practices
- 29:47Clinical RAG demo architecture
- 34:28Demo system features
- 37:30Summary and references
- 38:17Q&A on logs and PHI
related talks
27mOptimizing RAG Systems in Practice: Challenges and Proven Solutions
19mBuilding Production-Ready RAG Systems
02h 47mFrom naive to advanced RAG: the complete guide
59mThe Evolution of RAG
22mThe Power of RAG at Scale: A Year of Explosive Growth and Observab...
44mRAG for a Medical Company: Technical and Product Challenges AI & Data
01h 28mFrom RAG to Riches Building Enterprise Search Systems with LLMs
56mRAG: from dumb implementation to serious results