Operationalise Machine Learning and Generative AI solutions (Microsoft Certified MLOps Engineer)
About This Course
This 4-day course covers the end-to-end workflow for operationalising machine learning and generative artificial intelligence on Microsoft Azure. The first two days apply data engineering practices to Machine Learning Operations (MLOps): preparing and structuring data and selecting compute for training, training and comparing classification models with Azure Machine Learning, refining training with scripts and MLflow tracking, tuning hyperparameters with sweep jobs, and building data engineering pipelines that extract, transform and load data. Learners then use cloud computing application practices to plan an MLOps architecture, automate training runs with GitHub Actions, and publish and monitor models on batch and managed online endpoints.
The final two days extend cloud computing application and artificial intelligence application practices to Generative Artificial Intelligence Operations (GenAIOps): provisioning and configuring Microsoft Foundry cloud resources, managing and versioning artificial intelligence agent prompts with GitHub, testing and improving artificial intelligence agents through structured experiments, automating quality checks with cloud evaluators and GitHub Actions, monitoring cloud-hosted generative artificial intelligence workloads using Application Insights and distributed tracing, and applying artificial intelligence application techniques such as supervised fine-tuning, reinforcement fine-tuning and preference-based training. Every module includes guided hands-on exercises in a live Azure lab environment, and the course aligns to the skills measured in Microsoft Exam AI-300.
What You'll Learn
• Describe cloud computing application practices for an MLOps architecture in Azure, covering environment separation, source control, access security and GitHub Actions workflows that automate model training. (LO2)
• Provision and configure Microsoft Foundry cloud resources for generative artificial intelligence, with access permissions that secure the IT environment and factors informing language model selection. (LO3)
• Monitor cloud-hosted generative artificial intelligence workloads using Application Insights and distributed tracing to track latency, cost, errors and quality in production IT systems. (LO4)
• Explain how artificial intelligence application techniques such as supervised fine-tuning, reinforcement fine-tuning and preference-based training adapt models to business processes. (LO5)
• Apply artificial intelligence application skills to train, track and tune machine learning classification models in Azure Machine Learning using automated machine learning, MLflow and sweep jobs. (LO6)
• Apply data engineering practices to build and run pipelines that extract, transform and load data through reusable components in Azure Machine Learning, and publish a trained model to an online endpoint. (LO7)
• Develop and version artificial intelligence agent prompts in Microsoft Foundry using GitHub, and run structured experiments comparing prompt variants to strengthen artificial intelligence application. (LO8)
• Automate quality checks on artificial intelligence agent output using cloud evaluators in Microsoft Foundry, applying evaluation datasets and quality gates in an automated workflow. (LO9)
Entry Requirements
• Read, write, and speak English at WPL Level 4
• Manipulate numbers at WPN Level 4
• Comfortable reading/writing Python and working with scripts.
• Understanding of concepts such as training models, datasets, evaluation, inference, and model deployment.
• Basic DevOps knowledge in Source/version control
• Familiarity with Command-line tools
• Basic Azure familiarity