I'm a software professional and people leader, passionate about building AI‑powered healthcare platforms.
My recent work is a production‑grade Agentic Healthcare Platform that lets clinicians ask natural‑language questions and get clinical risk assessments backed by real FHIR data.
What it does:
A clinician types: “Is it safe to refill Kristle Mraz's amoxicillin?”
An AI agent searches a live FHIR server, fetches medications and allergies, runs a rule engine, and returns a structured risk assessment with an audit‑trail FHIR ID.
Why it’s agentic AI:
- The LLM (Claude) decides which tools to call and in what order.
- Ageit itself has no clinical logic – that lives in separate, testable services.
- The system is secure, FHIR‑compliant, and runs on GKE (but is cloud‑agnostic).
See it in action
📖 Full documentation & source code
🚀 Try it yourself – request an API key (see below)
I’ve deployed a live instance on Google Cloud. You can interact with it using the provided Python client.
- 📦 Install:
pip install fhir-clinical-client - 🔑 Request a trial API key (also feel free to message me on LinkedIn for a key)
- 📚 Then run the demo notebook or follow the README.
All code provided is available for review and critique, but my own customizations are fully proprietary and restricted based on included licensing terms. I have clearly indicated wherever I have used open‑source (E.g., Apache 2.0 for the core libraries).
Start here:
- Full architecture overview
- 85 unit & integration tests
- PR walkthroughs (each PR explains key decisions)
- Live demo patients already seeded
bhaskarcmu/bhaskarcmu is a ✨ special ✨ repository because its README.md (this file) appears on your GitHub profile.