Architecting for Innovation: Building AI Powered Apps and Platforms
summary
Kubernetes has 314,000 commits, 74,000 contributors and over 7,800 contributing organizations and the platform claim is that existing container platforms can run heavy AI stacks. Google highlights Cloud Run Application Canvas (Gemini-backed) to generate app scaffolds, GKE support for TPU v5p and A3 Mega nodes, and features like container image preloading that cut cold starts ~29x and GCS Fuse read cache for faster dataset reads; Gemini models (Ultra/Pro/Nano) and Vertex AI Model Garden enable side‑by‑side model selection.
AI-written summary · about
chapters (8)
- 0:00Introduction and agenda
- 1:02Kubernetes evolution and AI workloads
- 3:55Cloud Run and serverless AI apps
- 6:20Google Kubernetes Engine features
- 8:48AI workload optimization tools
- 11:42Vertex AI and model selection
- 12:02Demo: AI-assisted development
- 18:52New certifications and resources
related talks
27mHarnessing Kubernetes for Scalable AI/ML Workloads
20mOperating Frontier Intelligence at Scale
25mExtending Kubernetes for AI | Lessons Learned From Platform Engineering
23mThe State of Cloud Native: The Shift Towards AI
43mKeynote: Building AI in the Open: How the Cloud Native Community Shapes What Comes Next
14mKeynote: Building the Infrastructure that Powers the Future of AI
43mProblems and Solutions with AI Model Deployments in k8s
25mEmpowering AI-Driven Drug Discovery: Overcoming Challenges in Building a ML Platform on Kubernetes