Overview
What this challenge is about.
Simulate two sites with non-IID data splits (one site skews older, the other younger). Implement FedAvg using Flower (or PySyft). Run for at least 50 communication rounds; report per-site evaluation + global evaluation. Compare convergence and final accuracy to a centralized baseline. Add secure aggregation (lightweight version is fine for the prototype). Document the protocol's privacy properties precisely (what FL does and does not protect against). Write a 5-page technical spec the hospital IT teams will review for security sign-off.
The Brief
What you'll do, and what you'll demonstrate.
Prototype a federated-learning setup across two simulated hospital sites with documented privacy properties and a technical spec for IT review.
Earning criteria — what you'll demonstrate
- Implement FedAvg with non-IID client data
- Compare federated vs centralized training fairly
- Apply secure aggregation and document its threat model
- Communicate the limits of federated learning to a security audience
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Career paths this builds toward
Canonical rolesML Researcher
Federated-learning prototyping is the applied research work that healthtech and cross-institutional AI teams need on roadmap.
This challenge sharpens
- federated-learning
- fedavg
- distributed-systems
AI Safety Researcher
Documenting the threat model around federated learning and secure aggregation is exactly the AI safety research work that regulated AI teams need.
This challenge sharpens
- federated-learning
- secure-aggregation
- fedavg
AI Solutions Architect
Writing the technical spec that hospital IT signs off on is core AI solutions architecture work in healthcare and regulated industries.
This challenge sharpens
- federated-learning
- distributed-systems
- secure-aggregation