Tech lead in ML Ops team, making Waymo ML autonomous.

Keywords: Foundational Model Data Infra, MLOps, ML Pipeline, CI/CD, ML release, Model Metadata, Model life cycle management, Feature engineering, Feature Store, ETL Data Pipeline.

The Waymo ML Infrastructure team provides a set of tools and technologies to support and automate the lifecycle of the machine learning workflow, including feature and experiment management, model development, debugging & evaluation, deployment, and monitoring. These efforts have resulted in making machine learning more accessible to teams at Waymo, including Perception, Behavior Prediction, Planner, Routing, Maps and Research, ensuring greater degrees of consistency and repeatability, and addressing the “last mile” of getting models into production and managing them once they are in place. We work hand in hand with machine learning experts in all parts of the company as well as our collaborators across Alphabet. This is crucial to the development of Waymo Foundation Models for driving including evaluation, auto-labeling and distillation.

Marketplace Forecasting: Tech lead that collaborated with model researchers and data scientists to explore and productionize ML models to forecast demand and supply at global scale for dynamic pricing incentives.

▪ Designed and built a Tensorflow based application framework for Real-time forecasting, enabled faster model iteration by 8x and achieved model agnostic serving.

▪ Led Airport Demand project to generate trip forecasts to power ETR(Estimated Time to Request), productized 1st deep learning model for Real-time forecasting.

▪ Founding Eng for week-ahead forecasting system that runs GLM model to generate session/supply/trip/balance score forecasts, with holiday adjustment.

Uber Eats Courier Pricing:

▪ Owned batch pricing system that generates week-ahead VBF (Variable Base Fare) and Surge levels at global scale that powers courier baseline payment for Uber Eats.

▪ Led multiple efforts to build and iterate a Real-time fare system that calculates per-trip courier payment plus ML model-based incentive adjustments to reduce dispatch un-fulfillment and minimize cost.

Driver Growth:

▪ Built a document management service to store and process all kinds of driver documents during the onboarding phase, covering all 600+ Uber cities.

▪ Designed a web-based Appointment system for Ops to manage the driver's waiting queue during on-site inquiry and vehicle inspection.

• Built streaming pipeline to extract data stream from Yahoo mail to serve personalized data cards like hotel, rental car, package, flight, etc.

• Built real-time flight push notification scheduler system in a fast-paced three engineer scrum team, serving Yahoo mobile apps.

Participated in developing a big-data driven push notification system for gamers.

• Built an automated regression testing framework for mobile marketing web tool using Selenium

• Implemented Oozie workflow reconfigure tool for server partition expanding

• TA of Graduate Level Course: Open Source Tools (CSCI-GA 2246-001)