Microsoft - AI-900: Microsoft Certified: Azure AI Fundamentals

Sample Questions

Explanation:

Azure AI Speech service offers advanced speech to text capabilities. This feature supports both real-time and batch transcription, providing versatile solutions for converting audio streams into text.

The speech to text service offers the following core features:

• Real-time transcription: Instant transcription with intermediate results for live audio inputs.
• Fast transcription: Fastest synchronous output for situations with predictable latency.
• Batch transcription: Efficient processing for large volumes of prerecorded audio.
• Custom speech: Models with enhanced accuracy for specific domains and conditions.

Reference: What is speech to text?

Explanation:

Azure AI Foundry Models is your one-stop destination for discovering, evaluating, and deploying powerful AI models—whether you're building a custom copilot, building an agent, enhancing an existing application, or exploring new AI capabilities.

With Foundry Models, you can:

• Explore a rich catalog of cutting-edge models from Microsoft, OpenAI, DeepSeek, Hugging Face, Meta, and more.
• Compare and evaluate models side-by-side using real-world tasks and your own data.
• Deploy with confidence, thanks to built-in tools for fine-tuning, observability, and responsible AI.
• Choose your path—bring your own model, use a hosted one, or integrate seamlessly with Azure services.
• Whether you're a developer, data scientist, or enterprise architect, Foundry Models gives you the flexibility and control to build AI solutions that scale—securely, responsibly, and fast.

Azure AI Foundry offers a comprehensive catalog of AI models. There are over 1900+ models ranging from Foundation Models, Reasoning Models, Small Language Models, Multimodal Models, Domain Specific Models, Industry Models and more.

Our catalog is organized into two main categories:

• Models sold directly by Azure
• Models from Partners and Community

Reference: Explore Azure AI Foundry Models

Explanation:

The Azure AI Foundry playgrounds provide ready-to-use environments with all the necessary tools and features preinstalled, so you don't need to set up projects, manage dependencies, or solve compatibility issues. The playgrounds can accelerate developer velocity by validating API behavior, going quicker to code, reducing cost of experimentation and time to ship, accelerating integration, optimizing prompts, and more.

Playgrounds also provide clarity quickly when you have questions, by providing answers in seconds—rather than hours—and allowing you to test and validate ideas before you commit to building at scale. For example, the playgrounds are ideal for quickly answering questions like:

• What's the minimal prompt I need to get the output I want?
• Will this logic work before I write a full integration?
• How does latency or token usage change with different configurations?
• What model provides the best price-to-performance ratio before I evolve it into an agent?

Reference: Azure AI Foundry Playgrounds

Explanation:

Certain models in the Azure AI Foundry model catalog can be deployed as a serverless API deployment with pay-as-you-go billing. This kind of deployment provides a way to consume models as an API without hosting them on your subscription, while keeping the enterprise security and compliance that organizations need.

Each deployment has a rate limit of 200,000 tokens per minute and 1,000 API requests per minute. Microsoft currently limits serverless API deployments to one deployment per model per project.

References:

Deployment overview for Azure AI Foundry Models

Deploy models as serverless API deployments

Explanation:

Microsoft created a Responsible AI Standard, a framework for building AI systems based on six principles: fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability. These principles are the foundation of a responsible and trustworthy approach to AI, especially as intelligent technology becomes more common in everyday products and services.

Reliability & safety principle requires that AI systems must operate reliably, safely, and consistently. They should function as designed, respond safely to unexpected conditions, and resist harmful manipulation. Their behavior and ability to handle different conditions reflect the range of situations developers anticipated during design and testing.

This goal requires continuously assessing the system’s performance and safety after deployment to catch and correct issues or in other words, ongoing monitoring, feedback, and evaluation.

References:

What is Responsible AI?

Microsoft Responsible AI Standard, v2 (PDF)