Getting Started - AI Factory 101

Getting Started - AI Factory 101#

Take your AI Factory skills to the next level — master complex Assistants, advanced data pipelines, and enterprise-grade model serving.

Who Should Take This Path#

This intermediate learning path is designed for:

  • Application Developers integrating AI features into production systems

  • Data Engineers building robust pipelines for AI-driven Knowledge Bases

  • AI & Platform Engineers implementing model serving and observability

  • Enterprise Teams deploying Sovereign AI workloads at scale

Prerequisites: Completion of Getting Started - AI Factory 101 or equivalent experience with basic AI Factory concepts.

Learning Outcomes#

Upon completing this path, you will master:

OK Advanced AI Orchestration — Build multi-step Assistants with complex personas, persistent memory, and tool integrations

OK Enterprise Data Pipelines — Design Hybrid Knowledge Bases with multi-source RAG, metadata filtering, and hybrid search

OK Production Model Serving — Deploy and optimize GPU-powered models using KServe with performance tuning

OK Operational Excellence — Implement comprehensive observability, monitoring, and governance for AI-driven applications

Estimated Total Time: 2-3 hours | Difficulty: Intermediate

Learning Path Modules#

1. Advanced Assistant & Structure Design#

  • Master sophisticated AI orchestration patterns*

What You’ll Learn:

  • Design Assistants with nuanced personas and contextual memory

  • Build reliable multi-step Structures with error handling

  • Integrate external Tools and data sources securely

Time Investment: 30-45 minutes

Modules:#

  1. ** Assistants Reference Manual ** — Deep dive into assistant architecture, persona design, and memory management

  2. ** Structures Reference Manual ** — Master workflow orchestration with branching logic and validation patterns

  3. ** Create an Assistant in Gen AI Builder ** — Hands-on: Build a production assistant with integrated tools and knowledge bases

  4. ** Create a Structure in Gen AI Builder ** — Hands-on: Implement complex workflows with conditional logic and error recovery

  5. ** Create a Tool in Gen AI Builder ** — Hands-on: Wrap internal APIs as governed, secure AI tools

2. Data Engineering & Hybrid Knowledge Bases#

  • Build enterprise-grade data pipelines for AI*

What You’ll Learn:

  • Architect Hybrid Knowledge Bases for optimal performance and cost

  • Implement advanced RAG patterns with multiple data sources

  • Configure intelligent search with metadata filtering and ranking

Time Investment: 30-45 minutes

Modules:#

  1. ** Best practices for Hybrid Knowledge Bases ** — Proven patterns for balancing data freshness, coverage, and operational costs

  2. ** Manage Knowledge Bases in Gen AI Builder ** — Operational workflows: re-embedding strategies, source lifecycle management, and access controls

  3. ** Create a Retriever in Gen AI Builder ** — Hands-on: Fine-tune search behavior, implement advanced filtering, and optimize ranking algorithms

3. Model Serving with KServe#

  • Deploy and scale GPU-powered AI models*

What You’ll Learn:

  • Deploy high-performance models on GPU infrastructure

  • Optimize resource allocation and runtime configurations

  • Manage the complete model deployment lifecycle

Time Investment: 45-60 minutes

Modules:#

  1. ** Model Serving Concepts ** — Understand KServe architecture and deployment patterns

  2. ** How-To Configure a ClusterServingRuntime ** — Define optimized runtimes for model families with shared configurations

  3. ** How-To Deploy an NVIDIA NIM Microservices available on build.nvidia.comcontainer with KServe ** — Hands-on: Leverage NVIDIA NIM for accelerated inference performance

  4. ** How-To Update GPU Resources for an InferenceService ** — Hands-on: Right-size GPU and CPU resources for optimal cost-performance

  5. ** How-To Verify InferenceServices and GPU usage ** — Validate endpoints and monitor GPU utilization through Hybrid Manager

4. Observability & Production Monitoring#

  • Ensure reliable, monitored AI operations*

What You’ll Learn:

  • Implement comprehensive observability for AI pipelines

  • Monitor model performance, resource usage, and system health

  • Establish production readiness and SLA compliance

Time Investment: 20-30 minutes

Modules:#

  1. ** Observability for Model Serving ** — Configure metrics, logging, and alerting for models and data pipelines

  2. ** How-To Monitor deployed models with KServe ** — Track endpoint health, latency, and throughput in production

  3. ** Observability for Model Serving ** — Leverage platform dashboards and automated alerting systems

Continue Your Learning Journey#

What’s Next?#

Ready for Advanced Patterns? Progress to ** Advanced AI Factory Usage and Extensibility - AI Factory 301 ** to master:

  • Multi-agent orchestration and complex workflows

  • Advanced embedding pipelines and vector optimization

  • Enterprise governance and compliance frameworks

  • Scaling patterns for high-volume AI applications

Additional Resources#

Deepen Your Understanding:

Ready to Build Enterprise AI?#

Complete this 201 path to gain the expertise needed for deploying production-grade Sovereign AI applications with full control over your models, data pipelines, and operational infrastructure using EDB Postgres AI.

Start your first module: Assistants Reference Manual

Contents