Build a Secure Multi-Party Computation Demo for Cross-Bank Fraud Detection
Overview
What this challenge is about.
Pick MP-SPDZ as the MPC framework. Build a 4-party demo computing two protocols: (1) private set intersection of suspicious-account IDs across banks, (2) joint aggregate (count + sum) of fraud markers per merchant category. Run on a 4-VM local cluster. Benchmark wall-clock and bandwidth for set sizes 10K / 100K / 1M per party. Honestly identify where the cryptographic cost dominates (likely the 1M-row joint aggregate). Deliver source, the docker-compose 4-party setup, benchmark Jupyter notebook, and a 7-page memo recommending production / not, with operational concerns (key-management, semi-honest vs malicious threat model, regulator readiness).
The Brief
What you'll do, and what you'll demonstrate.
Build a 4-party MPC demo for joint fraud analytics and honestly assess production viability against a target of 1M rows per party.
Earning criteria — what you'll demonstrate
- Understand the MPC threat model (semi-honest vs malicious)
- Implement private set intersection and aggregate sums using MP-SPDZ
- Benchmark MPC protocols across realistic input sizes
- Communicate cryptographic guarantees to non-cryptographer banking stakeholders
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 mappings coming soon.