AI Factory How-To Guides (Hybrid Manager)

Use these How-To Guides to deploy and operate AI Factory capabilities inside your Hybrid Manager (HCP) environment.

Hybrid Manager runs AI Factory as an integrated AI workload inside your HCP Kubernetes project. While many core concepts are shared with the AI Factory Hub , there are important Hybrid Manager-specific patterns and steps to follow when:

• Setting up Model Serving infrastructure

• Deploying and verifying AI models (InferenceServices)

• Managing GPU resources for model serving

• Using AI models in Gen AI applications and database extensions

• Managing project- and user-scope permissions for AI Factory

Learn more

For conceptual background, see:

• AI Factory in Hybrid Manager

• AI Factory Hub — Full Concepts

• Setup GPU resources in Hybrid Manager
  • Goal
  • Estimated time
  • What you accomplish
  • What this unlocks
  • Prerequisites
  • Provision GPU nodes
  • Label and taint GPU nodes
  • Deploy the NVIDIA device plugin
  • Store NVIDIA API key as Kubernetes secret
  • Related concepts
  • Next steps
• Verify InferenceServices and GPU Usage in Hybrid Manager
  • Verify InferenceServices and GPU Usage in Hybrid Manager
  • Goal
  • Estimated time
  • What you accomplish
  • Prerequisites
  • Verify InferenceServices Status
  • Confirm GPU Resource Usage
  • Troubleshoot Common Issues
  • Related concepts
  • Next steps