Uber details closed-loop evals for a multimodal food agent at scale
AI Engineer · youtube · 2026-07-25
Uber explains how it built closed-loop evaluations for a multimodal food-enhancement agent used to improve dish photos for Uber Eats merchants.
The talk focuses on the hard parts: keeping edits faithful to the original dish, preserving each merchant’s branding and packaging, avoiding marketplace homogenization, and building evals without an existing playbook. It covers reward hacking, the mix of offline and online signals, and how Uber balanced creativity with strict safety guardrails at scale.
More from coding & agent
- Opus 5 plus Unity CLI feels like a cheat code for vibe-coding games — chongdashu · 2026-07-25
- OpenRouter launches beta classifiers to tag and visualize model routing — AccBalanced · 2026-07-25
- Codex workflow tip: draw the flow first, then ask it to implement it — alex_frantic · 2026-07-25
- Microsoft’s OpenForgeRL trains agents inside the same harnesses they use in production — dair_ai · 2026-07-25
- Paper argues production agents fail from context overload, not reasoning — omarsar0 · 2026-07-25
- Gemini CLI fixes Windows diff views by normalizing CRLF to LF in a2a-server — luisfelipe-alt · 2026-07-25