Advanced AI Factory Usage and Extensibility - AI Factory 301 ============================================================ Who this is for --------------- - Advanced developers building **AI-first applications** and agents - AI OPs teams managing **large-scale model serving** - Platform engineers building **extensions and integrations** with AI Factory - Hybrid Manager users building **enterprise-grade Sovereign AI** solutions ``---`` General Goals ------------- By completing this path, you will: - Build complex **agentic Assistants** with Tool calling and advanced RAG pipelines - Extend **Model Serving** with custom runtimes, Transformers, and Explainers - Automate and orchestrate **AI Factory pipelines** with APIs - Design **advanced observability and governance** for production AI - Integrate AI Factory with **enterprise systems** and existing AI pipelines ``---`` Modules by Focus Area ===================== ``---`` 1. Agentic Assistants and Advanced Tools ---------------------------------------- **Goals:** - Build Assistants that reason and act with multi-step Tool usage - Implement advanced Tools for external API integration - Chain Tools and manage context flow **Estimated Time:** ~45–60 min **Modules:** 1. :ref:`Assistants Reference Manual ` — Architect assistants that reason, retrieve, and act across tools. 2. :ref:`Tools Reference Manual ` — Design governed actions that call internal or external systems. 3. :ref:`Create a Tool in Gen AI Builder ` — Implement robust, reusable tool integrations for agent workflows. 4. :ref:`Structures Reference Manual ` — Build multi-step pipelines and agent behaviors. ``---`` 2. Extending Model Serving -------------------------- **Goals:** - Create custom **ServingRuntime** definitions for specialized models - Add **Transformers** and **Explainers** to Model Serving pipelines - Understand advanced deployment patterns for KServe-based serving **Estimated Time:** ~45–60 min **Modules:** 1. :ref:`Model Serving Concepts ` 2. :ref:`How-To Configure a ClusterServingRuntime ` — Create custom runtimes for specialized models and settings. 3. :ref:`How-To Deploy AI Models from Model Library ` *(placeholder — expand with custom Transformers/Explainers patterns)* ``---`` 3. Observability and Governance at Scale ---------------------------------------- **Goals:** - Implement enterprise-grade **observability** for AI pipelines and Model Serving - Monitor **Tool usage** and pipeline performance - Build custom dashboards for hybrid AI workloads **Estimated Time:** ~30–45 min **Modules:** 1. :ref:`Observability for Model Serving ` — Metrics, logs, and traces for serving and pipelines. 2. :ref:`How-To Monitor deployed models with KServe ` — Health and performance of serving endpoints. 3. :ref:`Observability for Model Serving ` ``---`` 4. Automation and API Integration --------------------------------- **Goals:** - Use **Hybrid Manager and AI Factory APIs** to automate workflows - Integrate AI Factory pipelines into CI/CD and AI systems - Manage large-scale AI Factory deployments via API-driven control **Estimated Time:** ~20–30 min **Modules:** 1. `Hybrid Manager API Overview `_ 2. `Using the Hybrid Manager API `_ ``---`` Next steps ========== After completing this 301 Path: - You will be equipped to design **Sovereign AI applications** at enterprise scale. - You will understand advanced integration patterns with AI Factory, Hybrid Manager, and existing AI pipelines. - You will be ready to contribute advanced workflows and reusable components (Tools, Structures, Transformers) to your AI Factory ecosystem. ``---`` Related learning resources ========================== - :ref:`Gen AI Builder Concepts (Assistants, Knowledge, Tools, Threads) ` - :ref:`Sovereign AI Explained ` - `Hybrid Manager: Using Gen AI Builder `_ - :ref:`Structures Reference Manual ` - :ref:`Model Serving Concepts ` ``---`` By mastering this 301 Path, you will be ready to drive **advanced AI innovation** across your organization — fully under your control — using **EDB PG AI** and **Hybrid Manager**. ``---``