AIA 100: Metadata-Driven AI Agents
Design AI agents whose behavior is driven by metadata instead of hard-coded logic. Build configurable, schema-driven agents that are easy to extend, govern, and scale.
AIA 100 is a hands-on, project-driven course delivered entirely inside a private GitHub repository. You'll learn the metadata-driven pattern for building AI agents — describing tools, prompts, routing, and policies as declarative configuration so agent behavior can change without rewriting code.
Through guided modules and real exercises, you'll architect agents that read their capabilities from metadata, wire up tools and data sources dynamically, and add guardrails, evaluation, and observability. Once you register and verify your email, you'll be invited directly to the course repository.
What you'll learn
- Understand the metadata-driven architecture for AI agents and when to use it
- Model agent tools, prompts, and routing as declarative configuration
- Build agents that load capabilities and policies dynamically from metadata
- Wire up tools, data sources, and APIs without hard-coding orchestration
- Add guardrails, evaluation, and observability to agent workflows
- Package and govern reusable, configuration-driven agent templates