I built a production‑grade ML system for real‑time transaction risk scoring using FastAPI, CI/CD, Docker deployment, and load‑tested performance (160 RPS, <150ms latency, 0% failure).
The repository has 1 star on GitHub.
Development started 2026-02.
The last update was in 2026-10.
Throughput: 160 requests/second.
Latency: <150 ms (average).
Failure Rate: 0% under sustained load.
Model ROC‑AUC Score: 0.93.
The API is deployed on Render.

