deploy-cluster

Automate SkyPilot-managed TPU cluster deployment on Google Kubernetes Engine.

5|Updated Feb 4, 2026

Search Tags:#machine learning #GKE #cluster deployment #TPU #SkyPilot

Ready to run

npx skills add https://github.com/primatrix/skills --skill deploy-cluster

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: deploy-cluster
Source: https://github.com/primatrix/skills/tree/main/plugins/exec-remote/skills/deploy-cluster
Command: npx skills add https://github.com/primatrix/skills --skill deploy-cluster

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires skypilot, gcloud, kubectl, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the deployment of SkyPilot-managed TPU clusters on Google Kubernetes Engine (GKE), ensuring that the required node pool exists for the requested TPU type, and simplifies the process of running multiple TPU types in parallel on the same GKE cluster.

Core Features & Use Cases

Automated Deployment: Automatically deploys SkyPilot-managed TPU clusters on GKE, creating a node pool if necessary.
Parallel Deployment: Supports running multiple TPU types in parallel on the same GKE cluster.
Use Case: For machine learning engineers who need to deploy different TPU types for different workloads, such as running unit tests on v6e-1 and e2e tests on v6e-4 in parallel on the same GKE cluster.

Quick Start

Deploy a TPU cluster on GKE by specifying the cluster name, TPU type, and zone.

Frequently Asked Questions about deploy-cluster

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I deploy SkyPilot-managed TPU clusters on GKE?▼

To deploy SkyPilot-managed TPU clusters on GKE, you specify the cluster name, TPU type, and zone. The deployment automates GKE cluster creation, node pool setup, and SkyPilot configuration for the requested TPU types.

Can I run multiple TPU types in parallel on the same GKE cluster?▼

Yes, you can run multiple TPU types in parallel on the same GKE cluster. This supports various machine learning workflows, allowing simultaneous deployment of different TPU configurations like v6e-1 and v6e-4.

Does SkyPilot automatically create GKE node pools for specific TPU types?▼

Yes, SkyPilot automatically creates the required GKE node pool for the requested TPU type if it does not already exist. This ensures the correct environment is provisioned for your machine learning workloads.

Do I need gcloud and kubectl to configure TPU clusters on GKE?▼

Yes, you need gcloud and kubectl installed to configure TPU clusters on GKE. These dependencies, along with skypilot, are required to interact with Google Kubernetes Engine and manage the cluster deployment.

What is the best way to run different TPU workloads on a single GKE cluster?▼

The best way to run different TPU workloads on a single GKE cluster is using automated SkyPilot deployment. It simplifies running unit tests on one TPU type and e2e tests on another in parallel without manual node pool configuration.

Why does my TPU deployment on GKE require SkyPilot configuration?▼

TPU deployment on GKE requires SkyPilot configuration because SkyPilot manages the cluster resources and schedules the machine learning workloads. It ensures the requested TPU types map correctly to the underlying GKE infrastructure.