Senior data science architect building enterprise-scale decision systems across credit strategy, forensic data integrity, regulatory remediation, risk modeling, and board-ready analytics.

End-to-End Python implementation of Christodoulides's (2026) interpretable forensic decision-support engine for institutional procurement integrity. Implements: graph-based supplier entity resolution, robust IQR standardization, Gaussian Mixture Model regime extraction, and a decomposable PHI composite score for prioritized audit triage.

Forensic-Integrated ML Pricing Engine for French MTPL insurance. Implements a two-stage frequency-severity ensemble (XGBoost, CatBoost, LightGBM) with Benford’s Law data validation, regulatory monotonic constraints, and SHAP-based interpretability.