Back to work
Healthcare · 2024

AI Health App

Graph RAG-powered maternity health platform delivering best-in-class Q&A accuracy

Patient satisfaction4.8/5
Medical accuracy94%
Documents indexed50K+

Mode

Project overview

We built a maternity health platform powered by a state-of-the-art Graph RAG system based on MIT research. The knowledge graph architecture delivers best-in-class accuracy on pre- and post-pregnancy questions — outperforming standard vector search on complex, multi-hop medical queries.

  • Graph RAG retrieval built on MIT research for multi-hop maternity Q&A
  • 50K+ medical documents indexed into a structured knowledge graph
  • 94% accuracy on complex pre- and post-pregnancy health questions

Discover

AI Health App

Phase 1 of 4

Discovery Phase

Mapped maternity health domain, reviewed MIT Graph RAG research papers, audited clinical data sources, defined accuracy benchmarks against existing solutions.

Deliverables

  • Domain audit report
  • Research paper review
  • Clinical data source map
  • Accuracy benchmark targets

Tech

OpenAILangChainPythonAWSPostgreSQLFHIR