Vancouver, British Columbia, Canada
UberPOOL Match Optimization Team(current) - Designed and deployed a dynamic programming approach to make dispatch decisions, using boosted trees to predict outcomes. Rolling out globally. - Derived theoretical bounds on efficiency gains from matching under various constraints, setting goals and directions for improving the matching algorithm. Uber Commute Team: - Founding data scientist: set team roadmaps, constructed data models, and designed experimentation and rollout plans. - Designed matching and dynamic pricing algorithms to maximize economic welfare. Other: - Applied clustering and classification algorithms to recommend pickup locations to riders. - Contributed heavily to internal R packages for experimentation and internal APIs. Taught in-house courses for data scientists in R and Shiny.