AI-300 Microsoft Certified Machine Learning Operations Engineer Associate Exam Prep

This exam prep course follows the official AI-300: Operationalizing Machine Learning and Generative AI Solutions skills measured, domain by domain:

Topic 1 Design and Implement an MLOps Infrastructure (15-20%)

• Create and manage resources in a Machine Learning workspace: workspaces, datastores, compute targets, identity and access management
• Create and manage assets: data assets, environments, components, and sharing assets across workspaces with registries
• Implement infrastructure as code: GitHub integration, Bicep and Azure CLI deployments, GitHub Actions provisioning workflows
• Restrict network access to Machine Learning workspaces and manage source control for ML projects with Git

Topic 2 Implement Machine Learning Model Lifecycle and Operations (25-30%)

• Orchestrate model training: MLflow experiment tracking, automated ML, notebooks, hyperparameter tuning, training scripts, distributed training and training pipelines
• Implement model registration and versioning: feature retrieval specifications, MLflow models, responsible AI evaluation and model lifecycle archiving
• Deploy models for production: real-time and batch endpoints, endpoint testing and troubleshooting, progressive rollout and safe rollback
• Monitor and maintain models in production: data drift, performance metrics, retraining and alert triggers

Topic 3 Design and Implement a GenAIOps Infrastructure (20-25%)

• Implement Foundry environments: resources and projects, managed identities and RBAC, private networking, Bicep and Azure CLI deployments
• Deploy and manage foundation models: serverless API and managed compute, model selection, versioning and deployment strategies, provisioned throughput units
• Implement prompt versioning and management: prompt design, prompt variants and comparison, version control for prompts with Git

Topic 4 Implement Generative AI Quality Assurance and Observability (10-15%)

• Configure evaluation for generative AI applications and agents: test datasets and data mapping, groundedness, relevance, coherence and fluency metrics
• Configure risk and safety evaluations and automated evaluation workflows with built-in and custom evaluators
• Implement observability: continuous monitoring in Foundry, latency and throughput, token and cost metrics, logging and tracing

Topic 5 Optimize Generative AI Systems and Model Performance (10-15%)

• Optimize RAG: similarity thresholds, chunk sizes, retrieval strategies, embedding model selection and hybrid search
• Evaluate and improve RAG with relevance metrics and A/B testing
• Implement advanced fine-tuning: fine-tuning methods, synthetic data, monitoring fine-tuned models and promoting them to production