Indian Institute of Technology Kharagpur
- MTech Thesis: Modelling of reinforced concrete elements under fire and predicting their capacity using various machine learning techniques. Objective was to model the instantaneous and transient creep strain due combined effect of high temperature and load using Finite Element Analysis in Abaqus. The data generated from this analysis was used for prediction using Huber regression, nuSVR, tree regression
- BTech Project: Configurational Thermostats for Extended Phase Systems. An attempt at sympletically solving the extended phase spaces used in Molecular dynamics by Trotter factorisation.
Piramal Finance
- NTC Underwriting model : Created XGBoost models that predict probability of default of customers leveraging geo-intelligence and payment gateway variables. Used clustering techniques on older data to increase data sample size for model developmen and PCA helped variable selection process
- Credit Risk Monitoring Dashboard : Analysis and reporting of various risk metrics for all the lending products.
- Anomaly detection framework : Developed the framework to identify anomalous branches using the attributes of the customers sourced.
- Model Monitoring Dashboard : Part of the team that monitors the effectiveness of all the underwriting models in production