Summary
Unit 7 of 7 · Module 5: Optimize generative AI model performance with Microsoft Foundry
Read this lesson
Packed and ready
That's Module 5 done. You learned several ways to get better results from a generative AI model in Microsoft Foundry, and how they fit together. You practiced prompt engineering, with system messages, few-shot examples, and model parameters. You saw when grounding a model with retrieval augmented generation pays off, and how it works. You learned when fine-tuning makes a model's behavior more consistent. And you compared all three, and saw when to combine them. The big idea: they're not rivals. Each one works on a different side of performance, so they fit together. Start with a good prompt to shape behavior, bring in RAG when accuracy depends on your own data, and reach for fine-tuning when style and format still won't stay put, however carefully you write the prompt.
The whole trip
The travel agency's best setup might use all three: a fine-tuned model that always sounds like the brand, RAG that answers from the real hotel catalog, and a prompt carrying the instructions and safety guardrails for each conversation. To dig deeper, Microsoft's documentation covers customizing large language models, prompt engineering techniques, system message design, RAG in Microsoft Foundry, fine-tuning and when to use it, and how to choose between RAG and fine-tuning. Next up, you'll learn how to build a responsible generative AI solution in Microsoft Foundry.
Facts from Microsoft's documentation: Prompt engineering techniques; System message design for Azure OpenAI; Retrieval augmented generation (RAG) and indexes in Microsoft Foundry; Customize a model with fine-tuning; Microsoft Foundry fine-tuning considerations (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.