Gen AI in Hybrid Manager

Hybrid Manager enables production-grade deployment of Gen AI capabilities through its AI Factory workload — delivering Sovereign AI solutions that integrate deeply with your Postgres data, AI pipelines, and enterprise architecture.

Hybrid Manager’s Kubernetes-native foundation, combined with AI Factory, allows you to build advanced agentic AI applications that operate on trusted data, run within governed infrastructure, and connect seamlessly to Postgres.

This page explains how Gen AI workloads operate in Hybrid Manager and how they fit within the broader EDB Postgres AI (PG AI) ecosystem.

For an architectural deep dive and full feature reference, start with the Gen AI Hub.

Architecture fit: PG AI and Hybrid Manager

Within the PG AI ecosystem:

• Postgres (AIDB / Vector Engine) serves as your trusted data layer.
• Pipelines prepare data for Knowledge Bases and AI applications.
• Model Serving (KServe) delivers high-performance inference.
• Gen AI capabilities (Agent Studio and Builder) orchestrate intelligent agents and conversational interfaces.
• Hybrid Manager provides the control plane and Kubernetes infrastructure to run all these workloads securely and observably.

Gen AI in Hybrid Manager is a first-class workload, enabling you to deploy:

• Assistants built via Agent Studio or Builder.
• Knowledge Bases that ground your agents in accurate data.
• Models served through your own infrastructure (Model Serving).
• Tools and Structures that extend agent capabilities and integrate with your systems.

All components run inside your Hybrid Manager project — supporting full governance, observability, and compliance.

Core components

Agent Studio

• A low-code / no-code interface for building and managing AI Assistants.
• Enables business teams, content owners, and developers to collaborate.
• Supports configuration of:
• LLM models (served via Model Serving)
• Knowledge Bases
• Rulesets and governance policies
• Tools for system integration
• Memory settings for context persistence

See: Agent Studio in Hybrid Manager.

Gen AI Builder

• A code-driven agent development environment, ideal for advanced AI developers.
• Supports creation of:
• Complex multi-agent workflows
• Tool-chaining and API orchestration
• Custom agent logic beyond basic chat
• Builders can leverage:
• Model Serving endpoints (KServe InferenceService)
• Knowledge Bases for Retrieval-Augmented Generation (RAG)
• Tools and Structures defined within the project
• Postgres-based data sources via AIDB and Vector Engine

See: Gen AI Builder in Hybrid Manager.

Solution patterns and use cases

Conversational knowledge assistants

• Build Assistants that answer domain-specific questions based on curated Knowledge Bases.
• Common for:
• Internal knowledge portals
• Compliance assistants
• Customer support AI agents

Enterprise copilots

• Develop Assistants that combine:
• AI reasoning (LLM)
• Live data from internal APIs (via Tools)
• Postgres-based transactional context
• Use cases:
• Sales and marketing copilots
• Financial advisors
• HR query agents

Process automation agents

• Automate business processes using AI:
• Categorization workflows
• Document summarization
• Data extraction and enrichment
• Powered by:
• Structures deployed in the AI Factory
• Pipelines feeding Knowledge Bases
• Model Serving for advanced inference

Hybrid search and RAG pipelines

• Build powerful hybrid search experiences by combining:
• Vector similarity search in Postgres (Vector Engine)
• Semantic search from Knowledge Bases
• Structured filters and metadata
• Key use cases:
• Enterprise document search
• Product catalog intelligence
• Compliance and audit search portals

Hybrid Manager-specific considerations

• All Gen AI workloads run inside the Hybrid Manager project’s Kubernetes cluster.
• Models served via Model Serving are secured with project-level controls.
• Knowledge Bases and Pipelines operate within the project’s storage and data governance boundaries.
• Observability for all Gen AI activity is integrated with Hybrid Manager monitoring:
• Inference metrics
• Assistant usage
• Pipeline health
• GPU resource utilization
• Hybrid Manager enables Sovereign AI:
• Your data stays in your project.
• Your models run on your infrastructure.
• All AI pipelines and outputs are auditable and observable.

See: Sovereign AI explained.

Learn more

• Gen AI Builder hub
• Agent Studio hub
• AI Factory Concepts
• Sovereign AI explained
• Hybrid Manager Model Serving
• AI Accelerator Knowledge Bases
• Hybrid Manager AI Factory overview