Diwesh Kumar

Developed classification model to predict the best customers basis transaction data, demographics, purchasing pattern etc to precisely target for acquisition by UHG. Worked on the B2B project to migrate process from Domo to Azure Data Bricks & Power BI. Developed sentiment analysis model to analyze the sentiment of respondents on various articles posted by UHG on twitter & Yammer. Developed Apriori algorithm based on Optum store data to create product bundles to increase AOV on Optum store site. Built the model to forecast the online enrollment of Medicare products for UHG at day level based on traffic, channels, and phone calls.

Analyze all the factors which can potentially contribute to customers becoming delinquent and provide a solution which can minimize the delinquency trend in customers. Perform Roll rate analysis to understand customer paying pattern/Dequincy behavior. Analysis of alternate channels like: CBOL/MBOL, ATM, OTP, E-mail etc. to increase RPC and reduce outbound calls to minimize the overall collection cost.

Designed work force optimization model for healthcare client. Project involved forecasting of incoming calls per day and on the basis of forecasted calls and other levers like SLA, AHT, # resources, overtime designed the work force manager for optimum allocation of resources. Developed classification model to predict dispute/ Invalid deduction 60 days in advance for Kellogg’s. Worked on to create dispute prediction application for the client. Lead multiple data science projects in the area of: forecasting, ML, optimization and helped client to productize and create app for them.

Worked on Linear Regression to find out the most important factors that influences customers purchase decisions in an Automotive Industry. Developed forecasting model to predict the sales of automobiles on the basis of historical sales data at make & trim level.

Worked on Fraud Investigation project to find out the overall efficiency of account managers who work on fraud related cases by analyzing the critical metrics information. Worked on Diners’ Club project to integrate 3 networks in SQL, SSRS and Excel which ultimately now is new network called Diners’ Club. Developed classification model to predict the customers who might become defaulters in future based on historical data.