• Evolution and Scale of Uber’s Delivery Search Platform 24 November 2025 / GlobalIntroduction Search is a primary discovery funnel for Uber Eats: a large share of orders start with people typing into the search bar to find stores, dishes, and grocery items. • Strong search directly translates into higher conversion, better basket quality, and faster time-to-order-especially for long-tail queries, new or seasonal items, and multilingual markets. • When search misses intent, people bounce or fall back to browsing. • When it understands intent, they find what they want in seconds. • Traditional search stacks begin with lexical matching, which is fast and effective when queries exactly match document text. • But real queries are challenging-synonyms (“soda” versus “soft drink”), typos (“mozzarela”), shorthand (“gf pizza”), language mix (“pan” meaning bread in Spanish, but a container for cooking in English), and context (“apple” the fruit vs the company).
Article Summaries:
- Uber Eats has upgraded its search engine from a lexical‑matching system to a large‑scale semantic search platform. The new architecture uses a two‑tower neural network built on the Qwen LLM, with query embeddings computed in real time and document embeddings generated offline. Training employs a Matryoshka Representation Learning (MRL) infoNCE loss, allowing multiple embedding dimensions for downstream tasks. The model is trained on hundreds of millions of data points using PyTorch, Hugging Face Transformers, Ray, and DeepSpeed ZeRO‑3 for distributed, mixed‑precision training. Production deployment includes an approximate nearest‑neighbor index, version control, and monitoring to support multilingual, long‑tail, and context‑rich queries across Uber Eats’ global markets.
Sources: