JEEVAN TARAKATURI

Data Science thought Partner to ICICI’s secured loan and ATM team and Digital activation. Built a linear Regression model for execution of cash forecasting model for ~15000 ATMs pan India expected to have savings of 117Cr monthly and a drop in cash out percent by 2% Went live in Hyderabad for 54 ATMs. Built a Random forest model and Spearheading the customer life-time value prediction project for the Auto Loan team for offering differential ROI to High NBC(Next Best Car) customers with recurring monthly revenue ~100Cr went live through Campaigns. Built C4M(Cloud For Marketing) Auto-Loan Model to identify elite customers who visited Auto loan RIB pages and dropped the offer while navigating through the pre-offered Auto loan using Google Cloud Platform(GCP) In live Overall conversion rate is 10-15%. Built a model for On-boarding Saving account Customers to make them Digital Txn Active by Recommending Products every week for 120Days(18 Weeks) went live through campaigns(Email, SMS, Notification) and the activation rate is (29%-43%). Analysing the dropout customers and pitching them through Campaigns of entire Auto Loan Journey. Building Models and Back testing of the models and Model Risk Validation.

Supported a team of 5 in Aggregating the list of comparable companies and analyzing future trends and growth for stock valuation to identify the trading opportunities. Worked on preparation of publicly available information to support the fundamental reports of stocks crypto for potential investment opportunities. Gained strong knowledge of risk control procedures, evaluating financial data and futures investment trends. Analysis of the Graphs and patterns and Customers impact and behavioural data in the Stock market analysis and Model Risk Validations.

Developed Churn Models to predict churn signals and recognize customers with an increased likelihood to discontinue SIM card services. Performed sensitivity analysis over models developed using classifiers like SVM (Support Vector Machine), Naive Bayes and Decision tree and concluded ’SVM’ as the best classifier which has predicted customer churn with 87.2% accuracy.