Senior Decision Scientist
Current• Designed a click-to-apply prediction model in python using LGBM for setting starting bids. Enhanced system accuracy by 51%• Developed a budget-splitting experimentation framework for auction bidding mechanism, transitioning from ‘wait-and-see’ to academically rigorous approach• Created a new canonical supply & demand framework from scratch for measuring marketplace health, being used by 6 separate teams• Programmed a bespoke experiment bucketing solution using linear sum assignment to fix test group allocation imbalances• Authored a strategic plan to transform our critical auto-bidding product from rules-based to AI-driven• Research on listing-dynamics resulted in new topline marketplace metric and various initiatives that improved application parity and speed