External validation of the wholesale PD, LGD and EAD models as a part of MRM Governance team for Bank of America - driving the model validation project starting from the creation of Initial assessment and Test plan creation to ultimately deploying the validation test on different statistical and portfolio aspects.

Responsible for building ALLL (Allowance for Loan and Lease Losses) models using Basel II framework for Commercial portfolio for a variety of products across different geographies:- Involved in validation of existing PD and LGD models to be able to incorporate them for ALLL reserves as the regulatory body is also forcing for the lifetime Expected Loss (EL) based reserves by 2018.- Developed 2 PD models and 3 LGD models using Linear Logistic regression, Fractional regression and Inflated Beta regression starting from the variable selection to the validation of the models.- PD models have been thoroughly validated using sophisticated statistical techniques by checking bad rate rank ordering, KS, Gini, Somer’s d, Area under ROC.- LGD models have been validated using LPCR (Loss Percent Capture Ratio), Expected Loss Shortfall and evaluating Exact and One-Notch matches in Confusion Matrix.- Also developed 2 challenger PD models using Cox’s Proportional Hazard (PH) method.- Model stability have been checked using PSI (Population Stability Index) and CSI (Characteristics Stability Index).- Mentored junior team members to complete the projects on time in a priority basis.- Have taken trainings on Linear and Logistic Regression, Time Series as well as SAS and Excel to help the junior team members to enhance their technical skills

Provided analytical solution to better manage credit risk in Automotive Finance domain – responsible for forecasting credit losses in Retail and Lease consumer business:- Developed and Implemented a triangulated version of Net-Flow and Vintage Loss Curve methods to predict credit losses for Consumer Retail business across 12 segments.- Forecasted different types of Consumer Lease losses for Automotive finance business. Lease business has Excess Mileage Wear and Tear (EMWT) losses and Principal deficiency losses other than payment due losses which needs special models other than the retail ones.- The Loss Forecasting models have been used in the CCAR process which needs stress testing of the models across 5 economic scenarios as per the regulatory requirement.- Have supported MIS reports to track important metrics in Retail and Lease businesses.