Behind the Ride: Uber's Cloud Architecture Powering Global Real-Time Matching

Abstract

This talk offers a deep dive into Uber’s real-time ride-matching platform, which handles millions of ride requests daily across thousands of cities. I will explore how Uber’s optimization engine reduced ride-matching times from several seconds to milliseconds, even during peak traffic surges.

The presentation will dissect how Uber’s transition from a monolithic system to a robust microservices architecture enabled significant improvements in global request latency, scalability, and resilience. Attendees will gain insights into how Uber processes terabytes of location data hourly using cloud-native solutions, edge computing, and AWS services to maintain high availability and reduce infrastructure costs per trip.

I’ll also cover the role of real-time market balancing through dynamic pricing algorithms and predictive analytics, showing how Uber forecasts demand and optimizes driver positioning to minimize rider wait times.

This session aims to share practical strategies for engineering high-performance, data-driven systems at scale—relevant to any organization navigating similar technical challenges.