AI at Lyft: From Data Foundations to Shipping Faster

​​Join Lyft Engineering during Toronto Tech Week 2026!

​AI only works when the foundation does. In this fast-paced lightning session, two Lyft practitioners cover both ends of the spectrum—followed by a panel discussion on lessons learned, trade-offs, and what's next.

Part 1 — Unifying Data and Metadata for Reliable AI
Shankar Vedaraman, Head of Corporate Data & Analytics

Most AI failures aren't model failures—they're data failures. Shankar shares how Lyft is converging a fragmented data estate into a metadata-driven platform, the prerequisite for trustworthy AI at enterprise scale.

Part 2 — Expanding Lyft to Europe with Cursor and Claude
Arshiya Sayyed, Engineering

A continent-scale Lyft + FreeNow integration should have taken a year of ramp-up. It didn't. Arshiya's team is building Roaming—letting Lyft riders hail rides across Europe—and she'll share how Cursor powered by Claude transformed how they plan, write, and review code.

We'll close with a panel discussion, connecting data foundations to developer velocity and exploring what it really takes to ship AI-powered systems faster in production. We'll share our perspectives on transforming Lyft's engineering towards AI-first development.