We present a production-oriented semantic retrieval system for Uber Eats that unifies retrieval across stores, dishes, and grocery/retail items. Our approach fine-tunes a Qwen2 two-tower base model using hundreds of millions of query-document interactions that were aggregated and anonymized pretraining. We train the model with a combination of InfoNCE on in-batch negatives and triplet-NCE loss on hard negatives, and we leverage Matryoshka Representation Learning (MRL) to serve multiple embedding sizes from a single model. Our system achieves substantial recall gains over a strong baseline across six markets and three verticals. This paper presents the end to end work including data curation, model architecture, large-scale training, and evaluation. We also share key insights and practical lessons for building a unified, multilingual, and multi-vertical retrieval system for consumer search.

翻译：我们为Uber Eats提出了一套面向生产环境的语义检索系统，实现了对商家、菜品及杂货/零售商品的统一检索。该方法基于数百亿条经过聚合与匿名化预处理的查询-文档交互数据，对Qwen2双塔基础模型进行微调。我们采用批内负样本的InfoNCE损失与难负样本的三元组NCE损失相结合的方式进行模型训练，并利用套娃表示学习技术从单一模型中生成多种维度的嵌入向量。该系统在六个市场与三个垂直领域中均较基线模型实现了显著的召回率提升。本文系统阐述了包括数据治理、模型架构、大规模训练及评估在内的完整技术方案。同时，我们分享了构建统一化、多语言、多垂直领域消费搜索检索系统的关键洞见与实践经验。