American Express (March 2024- August 2024)
  • Deployed Deep Learning models (LSTM, CNN) to predict cardmember spending behaviours, increasing offer redemption by 18% and driving $3M in incremental revenue
  • Led the end-to-end development of a proprietary Merchant Affinity Score (MAS) using Gradient Boosting algorithms (XGBoost, LightGBM) and customer transaction patterns, generating $1.5M in additional merchant spend
  • Partnered with data engineering team to optimize ETL processes in Merchant Offers Data Lake using Apache Spark, cutting pipeline execution time by 50%
  • Spearheaded deployment of a Scalable Machine Learning Pipeline (SMLE) on AWS Sage maker, reducing inference latency by 40% for real-time personalization of 20M+ card members
  • Detected and blocked over 10,000 fraudulent transactions using a robust Fraud Detection System (FDS) for DiMO Offers, leveraging Anomaly Detection and real-time streaming, while maintaining a false positive rate below 1 in 50 transactions
Northwell Health (January 2023- December 2023)
  • Performed Data Collection, Feature Engineering, Data modeling, and Visualization of budget and forecast data in python, utilizing K-Means Clustering to segment patient care cost patterns reducing ~$220K resource allocation discrepancies
  • Discovered High Utilizer Patients Report using SQL and Power BI reducing hospitalizations by 30% detecting ~200 patients at risk of readmission through the monitoring system
  • Scrutinized large Medicaid insurance claims datasets in RStudio, applying Generalized Linear Models (GLM) to develop optimized pricing and risk management strategies, Leading to a reduction in annual losses by 35%
Flexote IT Solutions (December 2020- August 2021)
  • Collaborated with cross-functional teams to develop and deploy InsightMax analytics suite boosting system performance and increasing user adoption by 10%
  • Leveraged SAS for customer data analysis and predictive modeling, contributing to the development of Salesforce Service Cloud features like automated customer support workflow led to 20% increase in customer satisfaction scores
  • Utilized Bayesian A/B testing, regression analysis, and machine learning models to optimize Subaru's Connected Vehicle features, leading to increase in user engagement across the STARLINK platform, impacting over ~50,000 users
IBM (June 2019- May 2020)
  • Created custom sales performance reports using IBM Cognos Analytics for Reliance Retail, enhancing insights into regional sales growth and top-performing products, optimizing sales forecasting accuracy by 15%
  • Designed SQL queries to extract defect rates and milestone adherence, for the IBM Watson AI using IBM DataStage and Db2, resulting in the resolution of ~50 critical defects and a 15-day refinement in meeting deadlines