Generative AI on Azure
This domain covers 15-20% of the AI-900 exam and focuses on generative AI concepts and Azure services.
What you'll learn
- Generative AI Fundamentals: LLMs, foundation models, tokens, hallucinations
- Azure OpenAI Service: Models (GPT-4, DALL-E, Whisper), deployments, endpoints
- Azure AI Foundry: Model catalog, hub/project model, prompt flow, evaluation
- Prompt Engineering: System messages, few-shot examples, temperature, grounding
- Responsible AI for GenAI: Content filtering, prompt injection, transparency, copyright
Azure Services Covered
| Service | Key Capabilities |
|---|---|
| Azure OpenAI Service | GPT-4o, GPT-4, GPT-3.5, DALL-E, Whisper, embeddings |
| Azure AI Foundry | Model catalog, prompt flow, evaluation, deployments |
Challenges
| Challenge | Topic | Focus |
|---|---|---|
| 19 | Generative AI Fundamentals | Core concepts, LLMs, scenarios |
| 20 | Azure OpenAI Service | Models, deployments, playground |
| 21 | Azure AI Foundry | Platform, model catalog, hub/project |
| 22 | Prompt Engineering Basics | Techniques, temperature, grounding |
| 23 | Responsible AI for Generative AI | Content filters, safety, transparency |