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：本文提出一种面向生产环境的语义检索系统，用于Uber Eats平台，实现对商家、菜品及杂货/零售商品的统一检索。该方法基于Qwen2双塔基础模型，利用数亿条经聚合与匿名化预处理的查询-文档交互数据进行微调。我们结合批次内负样本的InfoNCE损失与难负样本的三元组NCE损失进行模型训练，并采用套娃表示学习（MRL）技术实现单一模型支持多种嵌入维度。该系统在六个市场与三个垂直领域中，相较强基线模型均取得显著的召回率提升。本文系统阐述了包含数据治理、模型架构、大规模训练及评估在内的完整工作流程，同时分享了构建统一、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。