How Uber Eats Handles Billions of Daily Search Queries

Uber Eats has redesigned its search platform to manage billions of daily queries, accommodating nX merchant growth across multiple verticals. Challenges included expanding across discovery surfaces while maintaining low latency and high ranking quality. The search architecture employs a multi-stage pipeline for document ingestion, retrieval, ranking, and indexing, with strategies like latitude and hex sharding to balance load across time zones and handle high-density regions. Innovations such as ETA-based range indexing improve retrieval precision by prioritizing proximity and speed, resulting in significant latency and performance improvements.