Commercial/Consumer CCAR modeling; Loss Forecasting modeling; Commercial/Consumer CECL modeling

With a particular focus on modeling and stress testing for CCAR, DFAST and Basel

Work as part of the Consumer Risk Analytics team to develop and enhance portfolio models for reserves, capital and stress testing, as well as to enhance estimations of Basel II parameters for Consumer credit risk across different business lines. Perform reliability analysis and quality control of modeling data. Develop and maintain technical documentation for multiple projects, including model descriptions, mathematical derivations, data analysis, data quality control and processes. Provide support for validation, audit and supervisory reviews.

Joint modeling of longitudinal data and survival data with time varying coefficients

• Gained experience in consulting through active participation in short-term consulting projects• Provided recommendation report by explaining model behavior, identifying major significant sources of interest and carrying out scenario analyses• Fully exposed to SAS and other statistical software. Dealt with large data from real world problems• Obtained exceptional analytical, quantitative and problem-solving skills

•Gained experience in working with large data bases along with working knowledge of SAS and R•Provided a new frailty model which can distinguish transient effects (endogenous variables) and persistent effects (exogenous variables) in default related data•Applied generalized linear model with time varying coefficients in credit risk data to predict and forecast customer default probabilities•Implemented first hitting time model in survival analysis, and derived asymptotic results to make the model more flexible.

• Presented bias approximations for treatment effect in generalized linear models with the existence of treatment/covariates interactions through scenario analyses• Derived formulas for variance comparison between the true model and the misspecified model in the frame of generalized linear regression models• Experienced collaboration with members in a research group