Presentation: From Models to Agents: Building Context-Aware Consumer AI at Scale at DoorDash
Sudeep Das shares how DoorDash shifts from legacy one-shot predictions to an agentic recommendation platform. He discusses leveraging language-native consumer memory, RQ-VAE semantic IDs for catalog representation, and grounded search to dramatically boost relevance and conversion metrics. By Sudeep Das
Sudeep Das, Head of Machine Learning & AI for New Business Verticals at DoorDash, presented a talk on building context-aware consumer AI at scale. He explained how DoorDash transitioned from legacy one-shot predictions to an agentic recommendation platform. Das highlighted the use of language-native consumer memory, RQ-VAE semantic IDs for catalog representation, and grounded search to significantly improve relevance and conversion metrics.
As the leader of personalization, search, catalog intelligence, and decision-making systems for DoorDash's rapidly expanding consumer experiences in grocery, convenience, alcohol, and retail, Das shared insights from the company's production implementation of these AI systems. QCon AI, the event where he delivered the presentation, is a practitioner-led gathering focused on the engineering discipline required to scale these AI workloads safely, offering direct access to architectural playbooks and failure metrics from peer organizations.
Written by urgent.news from InfoQ's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
