Azure Data Lake Storage Gen2 Job Environments Azure Container Registry Data Prep Data Models Managed (Online/Batch) Endpoint Code Checkout Managed (Online/Batch) Endpoint GitHub Model Training Code Test, Data Check Model Training Model Evaluation Responsible AI Deploy to Stage Deploy to Stage Model Test Responsible AI container Gated approval Deploy to Prod Deploy to Prod Model Test ResponsibleAI Safe Rollout Compute Instance Compute Clusters build&push Training (pipeline) Deployment (pipeline) build&push job submit experiment log/metric Register model deploy automatically deploy automatically Event Grid (Model Trigger) (Code Trigger) model register Azure Machine Learning Python SDK/CLI Azure Machine Learning Python SDK/CLI Azure Machine Learning Components Azure Machine Learning Python SDK/CLI Prod Stage Dev Stage MLflow Model 形式を前提 とすることで、API デプロ イが簡素化 Registry により開発環境と 本番環境の分離が可能に