Capability Development / AI
Operationalize ML & GenAI Solutions (AI-300)
This redesigned AI-300 curriculum prepares learners to design, implement, automate, monitor and optimize production-grade machine learning and generative AI solutions on Azure. The program combines Azure Machine Learning, Microsoft Foundry, GitHub Actions, Azure CLI, Bicep, MLflow, responsible AI practices and observability techniques to build reliable AI systems aligned with modern MLOps and GenAIOps practices.
Program Overview
This redesigned AI-300 curriculum prepares learners to design, implement, automate, monitor and optimize production-grade machine learning and generative AI solutions on Azure. The program combines Azure Machine Learning, Microsoft Foundry, GitHub Actions, Azure CLI, Bicep, MLflow, responsible AI practices and observability techniques to build reliable AI systems aligned with modern MLOps and GenAIOps practices.
Audience Profile
For data scientists, AI engineers, machine learning engineers and DevOps professionals responsible for deploying, monitoring and maintaining AI solutions on Azure, including those preparing for the AI-300 certification or production AI operations roles.
Course Syllabus
Module 1: Design an MLOps Strategy
Topics
- MLOps operating model and business alignment
- Secure Azure Machine Learning architecture: workspaces, compute, registries
- Team collaboration across data science, engineering and DevOps
Labs / Practice
- Design an end-to-end MLOps architecture
Module 2: Experiment, Track and Tune Models
Topics
- Dataset preparation and automated ML experiments
- Run tracking, parameter logging and MLflow artifact management
- Hyperparameter tuning, sampling methods and sweep jobs
Labs / Practice
- Train, track and evaluate a classification model
- Run a sweep job and select the best model
Module 3: Build Pipelines and Automate Training
Topics
- Reusable Azure Machine Learning pipeline components
- Pipeline execution, monitoring and troubleshooting
- Automating model training with GitHub Actions and OpenID Connect
Labs / Practice
- Build and execute a reusable training pipeline
- Create a GitHub Actions workflow to trigger training
Module 4: Manage Environments, Releases and Deployment
Topics
- Environment versioning and MLflow model governance
- GitHub environments and controlled release workflows
- Managed online and batch endpoints, monitoring and rollback
Labs / Practice
- Register a model for controlled release
- Deploy, monitor and roll back a model
Module 5: Plan and Prepare GenAIOps Solutions
Topics
- Enterprise GenAIOps use cases and agent specifications
- Foundation model selection: capability, cost, latency, governance
- GenAIOps lifecycle: design, deployment, monitoring, evaluation
Labs / Practice
- Design agent specifications from business requirements
Module 6: Manage, Evaluate and Automate AI Agents
Topics
- Versioning prompts and agent configurations with GitHub
- Structured experiments: quality, safety, cost and latency metrics
- Automating AI evaluations with Microsoft Foundry and GitHub Actions
Labs / Practice
- Evaluate and compare agent versions with a scoring rubric
- Build an automated evaluation workflow
Module 7: Monitor, Trace and Optimize in Production
Topics
- Operational metrics: latency, cost, token usage and quality
- Distributed tracing for multi-step workflow debugging
- Production readiness: cost, safety and governance checkpoints
Labs / Practice
- Enable monitoring and tracing for a generative AI application