Prateek Srivastava

Manager - Credit Risk Analytics

HDFC Bank Ltd•  April 2022 - Present•  New Delhi, Delhi

Coordinated the implementation of internally created fraud triggers associated with Bank Statements to identify Simulated Bank Statements associated with Mule Accounts, as mandated by the RBI. Utilizing these triggers, we have also excluded these applications from Digital Credit Decisioning in the Business Loan Journey. Built and deployed a Machine Learning model for underwriter cost optimization, eliminating 12% of high-risk customers, achieving a 91% bad rate, and reducing underwriting costs by 15%. Spearheaded the integration of Bureau-like internal data sources, improving the Bureau Obligation Model’s predictive accuracy by 8%. Redesigned the bank statement-based income verification algorithm, increasing fraud detection accuracy by 19%. Collaborated with the Fraud Intelligence team to develop targeted checks for fraudulent salary credits, boosting detection rates by 15%. Introduced an OCR-based employment verification process for government employees, reducing manual verifications by 20% and generating Rs 33.45 Lakhs in projected annual profit.

Unit Manager - Precision Marketing

Bajaj Finance Ltd•  January 2021 - March 2022•  Pune, Maharashtra

A Geo-location-based statistical model was developed and implemented, accurate up to 14 meters on the ground. This innovation increased the precision of customer’s Geographical Boundary tagging by 27%, significantly improving targeted marketing efforts. Engineered an 87% accurate rule-based model for Personal Loan Propensity, enabling precise identification of high-propensity conversion journeys. Boosted conversion rates by 40% across personal loan and credit card channels, implementing Azure Event Hub API to enable near-real-time lead engagement for partner leads. Optimized UTM governance model via web scraping, reducing false attributions by 15% and improving tracking accuracy through parameter customization.

Business Analyst - Asset Management

EXL Services•  June 2018 - December 2020•  Bangalore, Karnataka

Designed and executed a lookalike modeling campaign to identify clients with similar portfolio profiles, generating $1 million in profit through enhanced client targeting and retention. Developed and validated a customer attrition model with 82% predictive accuracy, identifying high-risk accounts and implementing retention strategies that preserved investment capital. Built a cross-sell and up-sell recommendation model using ALS in PySpark, increasing cross-sell effectiveness by 30% and generating targeted data-driven fund recommendations.