Scaling up Lyft’s Growth via Causal ML and Bandits

Machine learning, causal inference, and reinforcement learning have played a key role in scaling up and personalizing Lyft decisions for growth, ranging from an ad exposure, a landing page for user onboarding, an email to be sent to a user, and a coupon/bonus to be dropped to rider/driver. In this talk, we’ll showcase a suite of growth engines powering Lyft’s rider, driver, and business growth, including user acquisition platform, communications platform, and incentives platform. We’ll discuss challenges in building real-world large-scale ML systems for business decisions, and share some learnings and best practices that can be ported to other industries beyond rideshare.

View the slides for this session