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.

Public portfolio of enterprise data strategy and governed decision systems spanning credit strategy, survival modeling, forensic data quality, reconciliation, regulatory remediation, validation, and executive decision support—built with SAS, SQL/PostgreSQL, and Python.