Presentation: From Models to Agents: Building Context-Aware Consumer AI at Scale at DoorDash

从模型到代理：在DoorDash大规模构建上下文感知的消费者AI

作者：Sudeep Das · InfoQ 原文

摘要：Sudeep Das分享了DoorDash如何从传统的一次性预测转向代理推荐平台，讨论了利用语言原生消费者记忆、RQ-VAE语义ID进行目录表示以及基于基础的搜索来显著提升相关性和转化指标。

Sudeep Das shares how DoorDash shifts from legacy one-shot predictions to an agentic recommendation platform.

Sudeep Das分享了DoorDash如何从传统的一次性预测转向代理推荐平台。

He discusses leveraging language-native consumer memory, RQ-VAE semantic IDs for catalog representation, and grounded search to dramatically boost relevance and conversion metrics.

他讨论了利用语言原生消费者记忆、用于目录表示的RQ-VAE语义ID以及基于基础的搜索来显著提升相关性和转化指标。

Sudeep Das currently serves as Head of Machine Learning & AI for New Business Verticals at DoorDash, where he leads personalization, search, catalog intelligence, and decision-making systems across rapidly expanding consumer experiences including grocery, convenience, alcohol, and retail.

Sudeep Das目前担任DoorDash新业务垂直领域的机器学习和AI负责人，领导个性化、搜索、目录智能和决策系统，涵盖快速增长的消费者体验，包括杂货、便利店、酒类和零售。

QCon AI is a practitioner-led event focused entirely on the engineering discipline required to scale these workloads safely.

QCon AI是一个由从业者主导的活动，完全专注于安全扩展这些工作负载所需的工程学科。

It provides direct access to the architectural playbooks and failure metrics that peer organizations use in production.

它直接提供同行组织在生产中使用的架构手册和故障指标。

阅读理解

1. What is the main shift described in the presentation?

2. Which of the following is NOT mentioned as a technique used by DoorDash?

3. What is the focus of QCon AI?

温故复习 →每日一句 →