Operationalize Machine Learning and Generative AI Solutions AI-300T00

Course Overview

Most organisations can build an AI proof of concept. Far fewer can run one reliably in production. AI-300T00 is a four day instructor led course that closes that gap, covering machine learning operations (MLOps) and generative AI operations (GenAIOps) on Azure.

Delivered as dedicated group training for your organisation, the course covers the full lifecycle of traditional machine learning models in Azure Machine Learning, then the deployment, evaluation, monitoring and optimisation of generative AI applications and agents in Microsoft Foundry. Your team works hands on with automation, CI/CD, infrastructure as code and observability using GitHub Actions, Azure CLI and Bicep. It also prepares them for the Microsoft Certified: Machine Learning Operations Engineer Associate certification.

Why should you do this course?

• Get AI out of the notebook and into production, with the automation and testing to keep it there.
• Give data science and DevOps a shared operating model, which is usually where AI projects stall inside a business.
• Prepare your engineers for the Microsoft Certified: Machine Learning Operations Engineer Associate certification, with the exam voucher included.
• Put evaluation, monitoring and tracing around generative AI, so quality problems surface before your customers find them.
• Standardise infrastructure with code, so environments stop drifting apart between experiment and production.

What will you learn?

• Run experiments, tune hyperparameters and build pipelines in Azure Machine Learning.
• Trigger machine learning jobs from GitHub Actions and work with feature based development.
• Manage environments and deploy models through an automated, repeatable process.
• Plan and prepare a GenAIOps solution, and manage prompts for agents in Microsoft Foundry with GitHub.
• Evaluate and optimise AI agents through structured experiments, and automate those evaluations.
• Monitor generative AI applications in production, and debug them with tracing.

Who is this course for?

• Machine learning engineers and data scientists taking models into production.
• DevOps and platform engineers who now own AI workloads alongside everything else.
• Teams where data science and engineering need one agreed way of shipping and monitoring AI.
• Engineers in your organisation preparing to sit the AI-300 exam.
• Organisations that have generative AI applications live and no reliable way to evaluate or monitor them.

What you need to know

This course sits at the intersection of data science and DevOps, and assumes some experience on both sides. Participants should be comfortable with Python, have a foundational understanding of machine learning concepts, and be familiar with practices such as source control, CI/CD and working from the command line. It is not an introduction to machine learning. Across four days your team covers the material Microsoft sets for the AI-300 certification path, applied to production style scenarios.

What's Next?

On completing the course your team will be prepared to sit Exam AI-300, the single requirement for the Microsoft Certified: Machine Learning Operations Engineer Associate certification. The exam runs for 120 minutes and is proctored through Pearson VUE, and an exam voucher is included with this course. View the certification details on Microsoft Learn.

Preparation resources for this exam, including the study guide and exam sandbox, are listed on Microsoft's certification page.

Microsoft role based certifications expire unless they are renewed. Renewal is free and is done through a short online assessment on Microsoft Learn.

Skills measured

• Design and implement an MLOps infrastructure
• Implement machine learning model lifecycle and operations
• Design and implement a GenAIOps infrastructure
• Implement generative AI quality assurance and observability
• Optimize generative AI systems and model performance

Course Structure

This course covers 13 modules across 2 topic areas:

1. Operationalising machine learning models (MLOps)

The full MLOps lifecycle for traditional machine learning: experimentation, pipeline automation, CI/CD, automated testing and deployment to production.

1. Experiment with Azure Machine Learning
2. Perform hyperparameter tuning with Azure Machine Learning
3. Run pipelines in Azure Machine Learning
4. Trigger Azure Machine Learning jobs with GitHub Actions
5. Trigger GitHub Actions with feature-based development
6. Work with environments in GitHub Actions
7. Deploy a model with GitHub Actions

2. Operationalising generative AI applications (GenAIOps)

The same discipline applied to generative AI: planning, prompt management, structured evaluation, automated testing, monitoring and tracing in production.

8. Plan and prepare a GenAIOps solution
9. Manage prompts for agents in Microsoft Foundry with GitHub
10. Evaluate and optimize AI agents through structured experiments
11. Automate AI evaluations with Microsoft Foundry and GitHub Actions
12. Monitor your generative AI application
13. Analyze and debug your generative AI app with tracing