3D RAG for Healthcare Documents

Context: Built at Cotiviti as part of the Generative AI Research Engineering track.

A 3D RAG system (semantic, lexical and graph) for parsing and retrieving complex healthcare documents:

• Cost-efficient pipeline: 1.5k docs for $0.5, end-to-end $10 for 1k docs with a $3M-$5M annual revenue opportunity
• All open source and fast execution
• Evaluation workflow for benchmarking RAG architectures using the MMLU methodology (multi-choice options)

Open source: agentic_rag - rag-consular - intro_rag
Press: Cotiviti