Skip to content

Managing AI Services in Infra Hub

AI Services are the fundamental building blocks of AI intelligence in AIDF. They are categorized into three types:

  • Data Services: Data processing, storage, and transformation
  • Model Services: ML/AI model capabilities (inference, fine-tuning)
  • Application Services: End-to-end AI applications

Service Visibility

Visibility Description
Public Available as a shared library across all projects in the workspace. Added to the common library. Can be licensed.
Private Available only within the specific project. Used for proofs-of-concept or proprietary services.

Adding an AI Service

  1. Switch to Config view
  2. Click AI Services → +Add
  3. Select visibility (Public/Private)
  4. Select the service ZIP file (must be uploaded to Azure Public Artifactory first)
  5. Click Submit — the Manifest Engine automatically builds and deploys the service
  6. A reviewer must approve the service before it can be used in pipelines

AI Service Management Options

Action Effect
Test Run a validation test with a sample payload before approving
Deploy Deploy a public service to the current project
Update Replace current version; old version is automatically undeployed
Undeploy Stop the service and release all allocated resources
Delete Completely remove the service and perform full resource cleanup