AI Engineering Stack: Assisted to Delegated
summary
Cloudflare reports 93% R&D adoption after wiring AI Gateway for model routing, Access for identity, Backstage as a service catalog, and Agents.md for agent context; an internal multi‑agent AI code reviewer enforces a machine‑readable engineering codex and blocks CI until AI approval; adoption spikes after Opus 4.5 and one‑command defaults; measurement relies on requests, tokens, model mix and a break‑glass workflow to track fixes and tune cost versus value.
AI-written summary · about
chapters (6)
- 0:00Cloudflare's AI engineering stack in practice
- 3:30AI Gateway, Access, Backstage, and Agents.md
- 4:51Getting to 93% R&D adoption
- 15:38Backstage as the system map for agents
- 21:06Measuring value beyond PR throughput
- 28:42AI code review as an objective gate
related talks
25mAI for Platform Engineering
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…
23mAI and APIs in 2026
30mAI APIs: Taming the Wild West with AI Gateways
30mAI Agents Unchained
28mAI for Everything After Code
19mThe New Application Layer
40mAI-Augmented DevOps with Platform Engineering
