Gen AI Builder in Hybrid Manager

Gen AI Builder in Hybrid Manager enables you to build advanced agentic AI applications — fully aligned with your Sovereign AI architecture — with complete control over data, models, and application behavior.

Builder applications run as containers in your project’s Kubernetes cluster and interact with:

  • KServe-powered Model Serving endpoints
  • Knowledge Bases built in your Hybrid Manager project
  • Tools and Structures for executing business logic and accessing external systems

For a deeper background, see the Gen AI Builder concepts page in the AI Factory hub and Gen AI Builder hub reference.


How Gen AI Builder works in Hybrid Manager

  • Gen AI Builder applications run in your project’s Kubernetes cluster — no external compute dependencies required.
  • Builders can call models deployed via Model Serving — see Model Serving.
  • Builders retrieve content from Knowledge Bases built in your Hybrid Manager project — see Knowledge Bases explained.
  • Builders invoke Tools implemented within your project — see Tools explained.
  • Builders manage Threads to enable stateful, auditable conversations — see Threads explained.
  • Builders can leverage Rulesets to enforce governance and behavioral consistency — see Rulesets explained.

Key Hybrid Manager considerations

  • Builder applications run on GPU-enabled infrastructure when required for high-performance inference — Update GPU Resources.
  • Builders access KServe-based Model Serving endpoints deployed within your Hybrid Manager project — Deploy AI Models.
  • Builders can access Knowledge Bases that are created within the project — see AI Accelerator Knowledge Base page
  • Builders operate entirely within your Hybrid Manager environment — supporting Sovereign AI principles:
  • Your data
  • Your models
  • Your AI pipelines
  • Full observability and control

See Sovereign AI explained.


Typical Gen AI Builder flow in Hybrid Manager

  1. Build or deploy ToolsHow-To: Create a Tool.
  2. Create and configure Knowledge BasesHow-To: Create a Knowledge Base.
  3. Deploy supporting models via Model Serving — How-To: Deploy a NIM container.
  4. Configure Rulesets as needed — How-To: Create a Ruleset.
  5. Build Assistants and/or agent workflows using Gen AI Builder — Gen AI Builder hub reference.
  6. Monitor Threads and tune behavior — How-To: View Threads.

Learn more



Could this page be better? Report a problem or suggest an addition!