About this project
FraudGuard is an end-to-end machine learning system for detecting fraudulent financial transactions. An XGBoost model was trained on millions of transactions, tracked with MLflow, tested with an automated pytest suite, containerized with Docker, wired into a GitHub Actions CI/CD pipeline, and deployed on AWS EC2 — this frontend calls that live API.
Model metrics
0.985
Recall
0.9996
ROC-AUC
0.710
F1 Score
0.556
Precision