Led Uber's Generative AI integrations platform, enabling secure, scalable generative AI across 30+ teams. Built integrations using Go, OpenAI APIs, Google Vertex AI, and custom private prediction to support 16M+ monthly queries at peak 25 QPS.

Machine Learning Engineer at Uber

July 1, 2021 - March 31, 2023

Redesigned driver earnings heatmap using Gaussian mixture deep models (PyTorch), improving earnings prediction accuracy and driver engagement by 18%. Optimized global incentive allocation with Ray + Spark hybrid pipelines, enabling 40× faster optimization and city-level scenario planning. Enhanced Upfront Fares & Trip Radius via distributed microservices; built a distributed simulation framework with Spark and agent-based modeling to accelerate policy validation. Implemented real-time surge pricing adjustments across 100+ metro areas. Embraced continuous feedback, transparency, and robust engineering practices.

Machine Learning Intern at Uber

June 1, 2018 - August 31, 2018

Researched model drift and data drift signals in historical payment logs; built visualization dashboards in Python to analyze customer behavior and risk indicators; supported product managers with data-driven insights.

Research Assistant at University of Michigan

November 1, 2016 - April 1, 2017

Collaborated with approximately 70 Michigan hospitals to review appendectomy datasets and improve surgical quality measures. Developed an index enabling meaningful cross-site comparisons; applied Spearman correlation, logistic regression, and ordinal logistic regression to assess surgical quality.

Research Assistant at Harvard University

October 1, 2017 - December 1, 2017

Combined a depression-related dataset for the US and China (2000–2012) using SQL; analyzed ~30 million cases via dataset imputation and regression; explored covariate correlations with linear, logistic, and ordinal logistic regression. Predicted depression distribution for 2013 using methods including LDA and hierarchical clustering; reduced RMSE with random forest.