Building an AI-Powered, Paved Road Platform with Cloud-Native OSS
summary
Intuit runs 3,000+ services across 3155 clusters and 16,000+ namespaces and builds an AI‑native paved‑road platform that maps an OAM‑inspired app spec to real Kubernetes resources. It wires Argo Rollouts + Numaflow for canary analysis and automatic rollback using anomaly scores, uses short‑window cluster components plus ML models for long‑window autoscaling recommendations (claiming ~25% developer velocity gain), and exposes ephemeral containers and Argo Workflows for one‑click diagnostics and heap/thread dumps, reducing MTTR and hiding Kubernetes complexity.
AI-written summary · about
chapters (11)
- 0:00Introduction and agenda
- 0:42Infrastructure overview
- 1:36Developer personas
- 2:51Addressing developer challenges
- 5:39Platform architecture pillars
- 7:07Runtime and autoscaling
- 11:04Network traffic management
- 11:58Interactive debugging
- 15:41On-demand debugging workflows
- 17:04Key takeaways
- 18:24Q&A session
related talks
05mKeynote: Intuit’s AI-Native Platform: Scaling to 8x Velocity for 8,000 Developers
26mBeyond Argo Rollouts: Boosting Developer Experience With Intelligen...
46mUnlocking The Tech Wonderland: Combining GitOps, Platforms, and AI
25mHow We Power the Largest AI Deployments on the Planet: Running Vir...
30mArgo CD Observability with AIOps - Detect Incident Fast
07mLightning Talk: Automating Map-building Pipelines for Safe Autonomous Driving...
07mHow Intuit Uses Argo, GitOps, and AI to Boost Developer Velocity
29mBold Move: Automating Away Developer Friction