LoopAI-103
Read this lesson

The job to fill

Every generative AI app sits on top of a foundation model, and picking the right one is your first big decision. Say you're building the support chatbot for a retail company, an outdoor-gear store with shoppers asking questions day and night. The bot has to get what customers mean, answer correctly, sound friendly, and never say anything unsafe. The catch: there are thousands of models to pick from. Which one do you choose? How do you know it's actually good at your job, and not just good on paper? And after launch, how do you measure it, and keep making it better? What you need is a repeatable process: discover models, compare them, deploy one, and check that it really meets your needs.

One platform, four floors

Microsoft Foundry walks you through that process, one step at a time. Step one, discover: the model catalog lists more than ten thousand models, from Microsoft, OpenAI, Anthropic, Meta, Hugging Face, and more. Step two, compare: line up your candidates on industry-standard benchmarks that score quality, safety, cost, and speed. Step three, once you've picked a winner, deploy it to an endpoint, the address your application calls. And step four, evaluate: run automated metrics and do some manual testing, to prove it clears your bar for quality and safety.

What you'll be able to do

In this module, you'll learn how to choose a model with confidence, which ways you can deploy it, and how to tell whether it's doing well. By the end, you'll search and filter the catalog, compare models side by side on benchmarks, deploy one and try it in the playground, and evaluate it both by hand and automatically, knowing which metrics fit which situation.

Facts from Microsoft's documentation: Microsoft Foundry Models overview; Model benchmarks and leaderboards in Microsoft Foundry; Understanding deployment types in Microsoft Foundry Models; Observability in generative AI (CC BY 4.0), rewritten in Loop's own words and examples. Narration, animation, and quiz by Loop. Microsoft, Azure, and Microsoft Foundry are trademarks of Microsoft. Loop is independent and not affiliated with or endorsed by Microsoft.