Find your next open-source project

Search by name, purpose, or keyword — or paste a GitHub repository URL
All skills

google-cloud-solution-guided-gke-ai-migration

Guides migrating an existing AI inference workload to self-hosted GKE inference with gcloud and kubectl.

  • Part of google/skills
  • 6.1k installs
  • DevOps & Self-hosting · Containers and clusters

When to use it

You already run inference on Cloud Run, the Gemini API, or a VM and need a GKE deployment plan with manifests and cutover steps.

For example

You say

Use google-cloud-solution-guided-gke-ai-migration to move my Cloud Run Llama 3 70B service to GKE on A100 GPUs.

You get
  • VRAM and hardware sizing
  • vLLM Deployment and Gateway manifests
  • Model staging Job manifest

How it works

  • Runs a four-phase flow: discovery, design, implementation, validation.
  • Defaults to vLLM behind an internal Gateway with Cloud Storage FUSE staging.
  • Writes progress to migration-state.md so work can resume later.