Learn how we developed a unified, global, Two-Tower Embedding model to replace 1,035 city-specific DeepMF models, decreasing training time by orders of magnitude and working at Uber scale with millions of eaters and stores and hundreds of millions of items

Denis Bolkovskis

Mental Health

Community Picks is a section on daily.dev where our community members share the most interesting and valuable content they've discovered online. From insightful articles to handy tools, every post is a gem curated by our dedicated coomunity. To contribute to Community Picks, you need to have at least 250 reputation points, ensuring that only active and trusted members can share their finds.

Community Picks

Uber has developed Two-Tower Embeddings (TTE) for its recommendation systems, using embeddings to improve scalability and efficiency. TTE provides personalized retrieval from a large pool of stores and can be used for final ranking in recommendation systems. The TTE model utilizes engagement data and localized relations to optimize training and inference. It offers feature extensibility and has been successful in improving performance and lowering costs at Uber. Challenges include model size, training time, and evaluation metrics.

Innovative Recommendation Applications Using Two Tower Embeddings at Uber

What are Embeddings and Two-Tower Embeddings (TTE)?