Years of Experience: 4. 10 Years
Design loss-forecasting frameworks (vintage, roll-rate/Markov, survival/hazard, GLM/GBM) at segment and portfolio levels.
Engineer features from bureau, internal behavioral, transaction, device, and alternative data with rigorous quality control.
Calibrate and backtest models; perform stability monitoring (PSI/CSI), discrimination (KS/AUC), and calibration tests.
Implement explainability (reason codes/SHAP), bias/fair-lending checks, and challenger/benchmark models.
Align models to accounting and capital frameworks (IFRS 9, CECL, or Basel concepts).
Drive portfolio spend growth through segment/persona targeting and campaign recommendations.