AI-300 Skills Measured

This objective covers the foundation for production machine learning: Azure Machine Learning workspaces, compute, environments, datastores, registries, networking, security, and collaboration. Candidates should understand how workspace design affects training, deployment, governance, and operations.

This objective focuses on training orchestration, experiment tracking, model registration, deployment, monitoring, retraining, and lifecycle decisions. Expect practical distinctions between experiment runs, registered models, deployment endpoints, production monitoring, and retirement.

This objective moves into Azure AI Foundry and Azure OpenAI operations. Study Foundry project environments, Model Catalog usage, Prompt Flow where applicable, agent and generative app configuration, identity, networking, deployment patterns, and operational setup.

This objective asks how teams evaluate and monitor generative AI applications and agents. Review evaluation datasets, groundedness, relevance, coherence, fluency, continuous monitoring, tracing, telemetry, token consumption, cost metrics, and resource usage.

This objective covers performance and quality improvements. Study RAG tuning, similarity thresholds, chunk sizes, retrieval strategy, hybrid search, fine-tuning decisions, model performance monitoring, cost tradeoffs, and when prompt or retrieval changes are safer than model customization.