Model Mastery brings together two complementary learning paths on Microsoft Foundry:
- Model-specific deep dives that explore partner models and their unique capabilities.
- Foundry-focused experiences that cover end-to-end agent development, evaluation, routing, and optimization on Foundry.
Each lab is modular and can be run as a short, focused exercise in 60–90 minutes or combined with related notebooks to form a longer 3–4 hour workshop. This makes the content flexible for live training, classroom delivery, or self-paced learning.
The content is intended for learners comfortable with Python, terminals, and cloud basics and is designed to be useful in both guided sessions and asynchronous study.
| # | Experience | Focus | Status | Start here |
|---|---|---|---|---|
| 1 | Anthropic on Foundry | Claude family deep dive | ✅ Ready | anthropic/ |
| 2 | Cohere on Foundry | Command A, Embed v4, Rerank v4 deep dive | ✅ Ready | cohere/README.md |
| 3 | Model router on Foundry Models | End-to-end optimization and routing on Foundry | ✅ Ready | foundry/model-router |
| 4 | Agent builder on Foundry Models | From Model Selection to Agent Optimization on Foundry | ✅ Ready | foundry/agent-builder |
Each learning path is fully self-contained. You can start with whichever model family you care about, or begin with a Foundry optimization track. Follow the setup steps in the folder you choose, and combine notebooks into a longer workshop whenever you want a deeper session.