Teach
Free starter lesson for the Foundry — gen-AI & agents module (Exam AI-901 Domain 2).
Foundry portal essentials
- Foundry provides a workspace to explore models, build AI apps, and manage deployments.
- Typical flow: open portal -> choose project/hub -> browse model catalog -> deploy model -> test in playground -> integrate endpoint.
- You usually manage keys, endpoints, and safety settings as part of deployment.
Models and deployments at a glance
- Model catalog helps compare available base models and capabilities.
- Deployment creates a callable endpoint for inference.
- Playground/testing validates prompts and behavior before application integration.
- Monitoring tracks usage, latency, and quality over time.
Optional employer-lab checklist (Azure portal + Foundry)
Use this optional checklist if your employer gives sandbox/subscription access:
- Sign in to Azure Portal.
- Search for "Azure AI Foundry" (or AI project resources) in your tenant.
- Open the assigned resource group and AI project/hub.
- Confirm you can access model catalog and deployment pages.
- Create or inspect one model deployment.
- Run a prompt test in playground and capture result notes.
- Review content safety settings and access controls before sharing endpoint details.
If step access fails, contact your organization admin for RBAC and subscription permissions.
Practice
Practice 1
In Foundry, where do you usually compare available base models?
Practice 2
Which step usually comes before calling a model from your app?
Practice 3
A deployment in Foundry primarily gives you:
Practice 4
Where should prompt behavior be validated before production integration?
Practice 5
If you cannot open Foundry resources in an employer lab, the likely first action is: