How Subaru Accelerated AI Model Development for Next-Generation EyeSight...
summary
Subaru builds an in‑cluster Harbor registry and uses Envoy Gateway in host mode plus MetalLB to cut 30 GB images pull time from about three hours to three minutes, migrates Helm scripts to GitOps with Helmfile and Argo CD to manage 25 application definitions, and models training as Argo Workflows to automate dependencies and pipelines, yielding a claimed 60× image pull speedup, reproducible deployments, and end‑to‑end CI for continuous AI model development.
AI-written summary · about
related talks
27mGen-AI at Scale: Simplifying Orchestration of Healthcare Applications Across Multi-Cluster Environme
36mNavigate the future of AI: Deploy generative models on Amazon EKS
28mInside BMW's AI Gateway: Self-Hosting Cost-Efficient AI at Enterprise Scale
12mBuilding Bulletproof AI Platforms
35mCase Study: AI-as-a-Service on Kubernetes at Scale and In Production
37mAGENTS OF S.E.A.L.E.D: AI Agentic Cybersecurity Framework
43mProblems and Solutions with AI Model Deployments in k8s
25mEdge-to-Core Machine Learning and Data Pipeline on k8s