GenAI Operations (GenAIOps) Workload Labs

The following hands-on exercises provide practical experience with GenAI Operations patterns and practices. You’ll learn to deploy infrastructure, manage prompts, implement evaluation workflows, and monitor production GenAI applications using Microsoft Foundry and Azure services.

Note: To complete the exercises, you’ll need an Azure subscription with sufficient permissions and quota to provision Azure AI services and deploy Microsoft Foundry workspaces. If you don’t have an Azure subscription, you can sign up for an Azure account with free credits for new users.

Quickstart exercises

Prerequisites for GenAIOps Labs

Set up your development environment with the required tools and accounts to complete all labs in this course.

Plan and prepare a GenAIOps solution

Deploy Microsoft Foundry resources and configure your development environment for building generative AI applications.

Develop prompt and agent versions

Create and deploy multiple versions of AI agents using prompt engineering and version management in Microsoft Foundry.

Design and optimize prompts

Use Git-based experimentation workflow to systematically test and evaluate prompt optimizations

Automated evaluation with cloud evaluators

Scale quality testing with automated cloud evaluators for systematic evaluation of AI agents

Monitor and trace your generative AI agent

Use Application Insights and distributed tracing to observe runtime performance and compare prompt versions of the Trail Guide Agent.

Optimize AI agents with fine-tuning

Analyze real agent quality problems, compare supervised fine-tuning, reinforcement fine-tuning, and direct preference optimization, and optimize simulated training.

Note: While you can complete these exercises on their own, they’re designed to complement modules on Microsoft Learn; in which you’ll find a deeper dive into some of the underlying concepts on which these exercises are based.