Harsharaj Subhash Narsale

Owned analytics for the Facebook Profile curation ecosystem, defining instrumentation strategy and analyzing user funnels to identify friction and inform product roadmap decisions. Partnered with Product, Engineering, and ML teams on AI-assisted profile editing, defining success metrics and evaluating feature adoption and engagement.

Built an XGBoost seller risk model in Amazon SageMaker that identified ~10K high-risk sellers, enabling 50+ investigators to prevent over 29M potentially fraudulent units from entering the fulfillment network. Led annual bad debt forecasting, building a Python engine using S-curve modeling and Monte Carlo simulation to streamline planning across payment methods and debt types.

Designed rule-based monitoring for credit card acquisition policies using PySpark and AWS operational datasets to identify policy violations and support regulatory compliance.

Built an autoregressive time-series model to forecast repeat booking rates with 90% accuracy, supporting flight demand, crew, and aircraft planning.

Performed pricing and campaign performance analysis to recommend optimal promotional discount strategies. Designed a dashboard to visualize customer price sensitivity and support pricing decisions.

Led a team of 10 Decision Scientists to improve campaign response rate by 1.5 pp through personalized targeting using customer segmentation (K-means) and purchase propensity modeling (Random Forest). Modeled route-level taxi and flight time distributions to forecast airline on-time performance with 97% accuracy using Monte Carlo simulations in Java.

Grade: GPA 3.98

Master of Science in Business Analytics from the University of Cincinnati, completed in August 2019 with a GPA of 3.98.

Grade: 81.8%

Bachelor of Engineering in Electronics and Telecommunication from Sardar Patel Institute of Technology, completed in May 2013 with 81.8%.