Transaction-monitoring anomaly detection (banking)
Anti-money-laundering anomaly detection for Italian banks: peer-group clustering, isolation forests, and behavioral feature engineering on transactional data.
I worked on anti-money-laundering software for Italian banks: anomaly detection for transaction monitoring, surfacing the customers and transactions worth an analyst’s attention among ordinary activity. The approach combined peer-group clustering, isolation forests and robust z-scores, on features engineered from transaction history, customer profile data and behavioral patterns. Data, results and client names stay under NDA.