Jul 2020 – Jun 2022 · 1 yr 11 mos

• Building horizontal Large scale Graph Learning capabilities at Mastercard for fraud detection, chargeback prediction, risk minimization.
• Leading an AI team at Mastercard to develop AI-powered risk assessment solutions for Issuers & Acquirers.
• Improved operational efficiency of multiple business verticals at Mastercard across multiple domains: NLP, Anomaly Detection, Classification, Time Series, Language Modelling, Optimization, etc.
• 15+ filed patents & 5+ research papers in Top tier ML conferences.
• Data challenge winner: CIKM 2020, KDD 2021, Kaggle (2 Silver, 1 Bronze)
• Research Interests: AI in finance, Graph Neural networks, Fairness AI

Mar 2019 – Jun 2020 · 1 yr 3 mos

Jul 2018 – Mar 2019 · 8 mos · Gurgaon, India

• Ad Serving on Home Page
• Created a Merchant Score for restaurants depicting its affinity of being clicked
• Built a weighted KNN Regression model for predicting the score of New Restaurants based on its content attributes
• Embedded cuisine features using Word2Vec and trained a regression model to get an improvement of 24% on CTR’s

Jun 2016 – Jul 2018 · 2 yrs 1 mo · Noida Area, India

• Media and Entertainment Industry:
• Delivered demand forecasting & movie scheduling solution for leading UK cinema chain.
• Devised a novel model for demand cannibalization of nearby shows using BLL & GLM in R;increasing scheduling of 3D format films by 30%.
• Optimized schedules using Dynamic Programming with screen heuristics using R and C integration for swift scheduling.
• Conceptualized demand spill-over approach using indifference theory increasing scheduling of big revenue films by 20%
• Gross Sales Model for Prospect Targeting(Fortune 20 BFSI Firm)
• Developed a segmentation based algorithm using correlation analysis & Text Matching for firmographic features.
• Started with a predictive feature set (∼600 features), performed correlation & bivariate analysis, multicollinearity & variable importance to get 35 best features for modeling
• Created an Ensemble of 3 ( 2 Regression, 1 Classification) GBM models to get a GINI of 90% on OOT data
• Travel Industry:
• High Precision Targeting for Customer Acquisition for an international cruise line company.
• Built stage wise analytical models to identify customers with high propensity of conversion over Direct/Electronic mail marketing
• Used various statistical parameters to validate stability and effectiveness of model.

May 2015 – Jul 2015 · 2 mos · Benguluru

• Cross Platform App Development
• Structured the inventory DB with MySQL; cached in Redis to develop an improved Ordering Logic for products
• Designed and implemented a fully functional Aging logic to terminate recrudescing tasks using PHP in HTML5 Apps
• Assimilated Google MapActivity using Android SDK in Eclipse IDE; appended markers for source & destination
• Fabricated a Location Listener for refreshing marker based on GPS provided LatLng values with automatic zoom