Skip to main content

Section

Selected Projects

Products I designed, built, and run end to end, from the backend and data layer through deployment and observability.

Project 1 post

Fraud Detection

MSc thesis on card fraud detection over 590,540 IEEE-CIS transactions: leakage-safe chronological evaluation, Optuna tuning, MLflow tracking, SHAP-driven feature reduction, and a FastAPI service that returns an explanation with every score. Graded 10/10.

Python LightGBM XGBoost CatBoost scikit-learn Optuna MLflow SHAP FastAPI

More builds

GitHub includes additional experiments, ML work, and implementation details beyond the highlighted projects here.

Open GitHub (opens in new tab)