Akash Gupta

- Validating regulatory Credit Risk models (PD, LGD, EAD) for the Retail and Wholesale portfolios of Tier-1 global Banks under IFRS-9 and CAR frameworks, ensuring accuracy, transparency, and compliance with supervisory standards. - Developed benchmark PD term structure and EAD models, leveraging segmented Decision Trees and Spline-based Logistic Regression frameworks, achieving a 5% AUC lift over existing models. - Designed and executed scenario stress tests for PD models under Macroeconomic Conditions.

- Led end-to-end development of Acquisition and Behavioral scorecards for Banks, NBFCs, and Fintechs, reducing delinquency risk while improving approval strategies. - Built a modular Automation framework for feature engineering (IV, WOE encoding, data processing), cutting compute time by 60% and reducing turnaround time in model development operations.

- Built loan and customer level Logistic Regression models to predict delinquency migration, applying IV-based and Stepwise feature selection with strict VIF controls to eliminate multicollinearity. Models achieved high stability and accuracy, delivering strong GINI/KS metrics with minimal PSI drift on out-of-time validation. - Conducted ongoing scorecard monitoring and validation, tracking GINI, KS, PSI, and CSI metrics to ensure portfolio stability and early identification of data drifts.

- Led a team of 3 analysts to build Expected Credit Loss (ECL) models for MSME Overdraft portfolios of a large Public Sector Bank, in line with IFRS-9 guidelines, directly supporting regulatory submissions. - Leveraged SQL, Python, and R for large-scale data mining and risk analytics, enabling accurate provisioning and enhancing client portfolio oversight.

- Developed Machine Learning classification models (Logistic Regression, XGBoost, Neural Networks) to optimize cross-sell targeting for automotive clients, increasing campaign conversion rates. - Designed SAS Macro libraries for WOE encoding and SMOTE sampling, improving efficiency and standardization of model development pipelines.

- Directed pre-production workflows for digital animation projects, aligning sound design with creative and technical objectives to deliver high-quality media outputs. Actively leveraged market research to optimize content strategy, project quality, and delivery efficiency.

Core subjects: Multivariate Statistics, Econometrics, Python Programming, Credit Risk Management, Machine Learning