AI Digital Twins: From YC Hackathon to Production with Apollo MCP
summary
A four‑hour hackathon prototype becomes a production system in a few months, reaching $100K ARR by connecting FHIR and legacy EHR APIs (Epic read/write) via Apollo MCP + GraphQL. The deployed stack integrates PubMed search and ingests ~100–200 papers per query, uses narrow specialty LLMs, Phonic voice models, graph neural nets for relationship reasoning, automates summaries, multilingual patient conversations and caregiver workflows — and is live with clinicians.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 3:09Early coding beginnings
- 9:45Healthcare data platform
- 11:53AI digital twin demo
- 15:46Healthcare as platform
- 21:26YC hackathon to MCP
- 27:10Automation in clinical AI
- 33:09Phone call demo Latino care
- 40:26Key takeaways
- 43:36Q&A
related talks
30mFrom Code to Clinic — Building an Internal Developer Platform for Clinical AI
59mArtificial intelligence in Healthcare
31mBuilding SwasthaX: Engineering a Production-Scale AI Platform for Healthcare
47mThe Future of Healthcare and AI
33mBridging large-scale FHIR to AI in production
27mPanel: APIs for AIs
36mAI in healthcare - what is hype and what is real?
31mAI For Good: Fighting Health Insurance with AI