Ronik Kumar Email & Phone Number

Noida, Uttar Pradesh, India

Gurgaon, Haryana, India

- Conduct CCAR Stress testing and forecasting based on different macroeconomic conditions.-Spearheaded the analysis and delivery of the model outputs of 9 cohort recovery and vintage GCLR models during the CCAR2020 Exercise.- Identified areas of enhancement and automated the production architecture in SAS which reduced model output delivery time by 40%.- Perform Quarterly monitoring to track the model performance and identify the root cause in case the model performance breaches the threshold.- Implemented age default cohort level fractional logit regression models to forecast potential losses and recoveries- Spearheaded the requirement for conducting CCAR2020 exercise through Autosys.-Created an efficient pipeline for obtaining model output from Autosys by creating JIL Scripts from scratch.

Gurgaon, Haryana, India

-Forecasted future cashflows for multiple credit card portfolios of one of the largest bank in USA and provided analytical insights into the trends observed of the key credit card cashflows metrics.- Formulated a methodology to incorporate seasonal variations in cashflow drivers by devising a strategy to generate seasonality factors.- Conducted root-cause analysis to understand the variance between the actuals and the forecasts and strategize the forecast methodology.- Optimized the budget allocation strategy of the marketing team by forecasting cashflows of future credit card offers and performing Scenario Analysis by weighing the profitability of these multiple scenarios of the future offers of the credit card portfolio.- Created a dashboard using the R shiny package in R for better visualization of the comparison between the forecast and historical actuals.

Greater Noida, Uttar Pradesh, India

Worked with one of the largest bank based in USA for the development of machine learning model to predict the true call reason behind the customer calling the “IVR customer care” unit of the client.- Set up a pipeline for migration of data from Teradata to Hadoop filesystem(HDFS) by leveraging Scoop.- Data profiling and generation of signals to capture customer behaviours and triggers for the IVR calling.- Leveraged Signal Hub to train and tune XGBoost Machine Learning model on the training data.- Validated the model performance using the ROC curve.

Integrated M.Sc., Physics, 8.00

Pcm, 91.00%