Pharma capstone: medical-information assistant
Solve a real pharma problem end to end, such as a medical-information assistant, with audit logs, safety review and monitoring.
The brief
Pharma companies answer thousands of medical-information questions from doctors and pharmacists. Build and deploy an assistant (or another pharma AI product you scope with your mentor) grounded in official labels, with audit logs, a safety review and monitoring. Good looks like: a deployed product, measured accuracy, a safety and compliance review, and a demo for a panel.
What you will build
- Scope the problem and success measures with your mentor
- Build the assistant grounded in official labels
- Add audit logs and safety rules
- Evaluate accuracy and safety on a test set
- Deploy with monitoring
- Present the results to a panel
Free, open data to build it with
- openFDA drug labels ↗ U.S. Food and Drug Administration
- ChEMBL ↗ EMBL-EBI
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.

