From Models to Agents: Building Context-Aware Consumer AI at Scale at DoorDash @infoq
From Models to Agents: Building Context-Aware Consumer AI at Scale at DoorDash @infoq
Uploaded August 2026 | Updated September 2026, 2 weeks ago
"LLMs aren’t going to understand what an embedding means. These are just numbers."

Sudeep Das (Head of ML/AI at DoorDash) breaks down why agentic recommendation systems are shifting from numerical expressions to language-native memory snippets.

Instead of relying on opaque vectors, passing semantic, text-based context allows downstream LLMs and agents to actually understand consumer behavior.

🔗 Watch the full presentation with transcript: bit.ly/3UW9Cko

#InfoQ #QConAI #MachineLearning #LLMs #AIAgents #VectorSearch #RecommendationSystems #DoorDashEngineering #DataEngineering #SoftwareArchitecture
From Models to Agents: Building Context-Aware Consumer AI at Scale at DoorDashHow Kraken Scales Serverless to Balance the National GridImproving Valkey with Madelyn OlsonOpen Source LLMs: Costly Myths & Real-World ScalingCloud Security Challenges in the AI Era - How Running Containers and Inference Weaken Your SystemAmazon DynamoDB Distributed Transactions at ScaleGenerally AI Episode 1: Large Language ModelsThe Latest in OpenJDK and JCP Expert Group: Insights with Simon RitterMake Quality Executable: Verification Gates for Agentic Engineering at ScaleAmazons 3000% Growth: The Hard Truths of Engineering at Scale20 years of InfoQ 🎂 #AlmirVuk #InfoQ20 #DevCommunityPrompt to Prod: Engineering an Autonomous SDLC at Scale
InfoQ |

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

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER