Major project
Banking & Fintech
Real-time fraud scoring service
Build a fraud model and run it as a low-latency scoring API with monitoring and drift alerts.
The brief
A payments company must score every transaction in milliseconds. Build the fraud model, then deploy it as a low-latency API with logging, versioned models and monitoring that raises an alert when transaction patterns drift. Good looks like: a model evaluated for imbalance, an API that meets a latency budget under load, a canary release of a new version, and drift alerts.
What you will build
- Build and evaluate the fraud model for extreme imbalance
- Register models and track experiments in MLflow
- Serve scores from a FastAPI service in Docker within a latency budget
- Load-test the service
- Release a new version as a canary, with rollback
- Monitor drift and alert the fraud team
Free, open data to build it with
- Credit Card Fraud Detection ↗ Kaggle (ULB)
- PaySim mobile money transactions ↗ Kaggle
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.