vocabulary: "1.0.0" info: provider: "Amazon SageMaker" description: "Unified taxonomy mapping operational and capability dimensions of the Amazon SageMaker machine learning platform APIs." created: "2026-04-19" modified: "2026-04-19" operational: apis: - namespace: amazon-sagemaker version: "2017-07-24" baseUrl: https://api.sagemaker.{region}.amazonaws.com status: active - namespace: amazon-sagemaker-runtime version: "2017-05-13" baseUrl: https://runtime.sagemaker.{region}.amazonaws.com status: active - namespace: amazon-sagemaker-feature-store version: "2020-07-01" baseUrl: https://featurestore-runtime.sagemaker.{region}.amazonaws.com status: active - namespace: amazon-sagemaker-metrics version: "2022-09-30" baseUrl: https://metrics.sagemaker.{region}.amazonaws.com status: active resources: - name: NotebookInstances api: amazon-sagemaker actions: [Create, Describe, List, Start, Stop, Delete, Update] - name: TrainingJobs api: amazon-sagemaker actions: [Create, Describe, List, Stop] - name: Models api: amazon-sagemaker actions: [Create, Describe, List, Delete] - name: Endpoints api: amazon-sagemaker actions: [Create, Describe, List, Delete, Update] - name: EndpointConfigs api: amazon-sagemaker actions: [Create, Describe, List, Delete] - name: Experiments api: amazon-sagemaker actions: [Create, Describe, List, Delete, Update] - name: Pipelines api: amazon-sagemaker actions: [Create, Describe, List, Delete, Start, Stop] - name: FeatureGroups api: amazon-sagemaker-feature-store actions: [Create, Describe, List, Delete, GetRecord, PutRecord, DeleteRecord] actions: - name: Create httpMethods: [POST] pattern: write - name: Describe httpMethods: [POST] pattern: read - name: List httpMethods: [POST] pattern: read - name: Delete httpMethods: [POST] pattern: destructive - name: Update httpMethods: [POST] pattern: write - name: Start httpMethods: [POST] pattern: write - name: Stop httpMethods: [POST] pattern: write - name: Invoke httpMethods: [POST] pattern: write schemas: core: - name: NotebookInstance description: ML compute instance with Jupyter notebook for ML exploration and development - name: TrainingJob description: A managed ML training job that trains a model on specified data - name: Model description: A trained ML model registered in SageMaker model registry - name: Endpoint description: A deployed model inference endpoint that accepts prediction requests data: - name: Tag description: Key-value metadata tag applied to SageMaker resources parameters: pagination: - name: NextToken description: Token for paginating through result sets - name: MaxResults description: Maximum number of results to return per page identifiers: - name: NotebookInstanceName description: Unique name identifying a notebook instance - name: TrainingJobName description: Unique name identifying a training job - name: ModelName description: Unique name identifying a model - name: EndpointName description: Unique name identifying an inference endpoint enums: notebookStatus: - Pending - InService - Stopping - Stopped - Failed - Deleting trainingStatus: - InProgress - Completed - Failed - Stopping - Stopped endpointStatus: - OutOfService - Creating - Updating - SystemUpdating - RollingBack - InService - Deleting - Failed instanceTypes: - ml.t3.medium - ml.t3.large - ml.m5.xlarge - ml.p3.2xlarge - ml.g4dn.xlarge authentication: schemes: - type: AWS Signature V4 description: AWS IAM-based request signing using access key and secret key apis: [all] capability: workflows: - name: ML Lifecycle Management file: capabilities/ml-lifecycle-management.yaml description: End-to-end ML lifecycle from notebook development through training, model registration, and endpoint deployment apisConsumed: [amazon-sagemaker] toolCount: 10 personas: [ML Engineer, Data Scientist] domains: [Machine Learning, MLOps] personas: - id: ml-engineer name: ML Engineer description: Engineers who build, train, and deploy machine learning models at scale workflows: [ML Lifecycle Management] - id: data-scientist name: Data Scientist description: Scientists who explore data and experiment with ML models using notebooks and experiments workflows: [ML Lifecycle Management] domains: - name: Machine Learning description: Core ML activities including model development, training, and inference resources: [NotebookInstances, TrainingJobs, Models, Endpoints] - name: MLOps description: Operational management of ML pipelines, monitoring, and model governance resources: [Pipelines, Experiments, Endpoints] - name: Feature Engineering description: Feature creation, storage, and retrieval for ML models resources: [FeatureGroups] namespaces: consumed: - amazon-sagemaker rest: - ml-lifecycle-api (port 8080) mcp: - ml-lifecycle-mcp (port 9090) binds: - name: AWS_ACCESS_KEY_ID workflows: [ML Lifecycle Management] - name: AWS_SECRET_ACCESS_KEY workflows: [ML Lifecycle Management] - name: AWS_REGION workflows: [ML Lifecycle Management] crossReference: - resource: NotebookInstances operations: [CreateNotebookInstance, DescribeNotebookInstance, ListNotebookInstances] workflows: [ML Lifecycle Management] personas: [ML Engineer, Data Scientist] - resource: TrainingJobs operations: [CreateTrainingJob, DescribeTrainingJob, ListTrainingJobs] workflows: [ML Lifecycle Management] personas: [ML Engineer, Data Scientist] - resource: Models operations: [CreateModel, DescribeModel, ListModels] workflows: [ML Lifecycle Management] personas: [ML Engineer] - resource: Endpoints operations: [CreateEndpoint, DescribeEndpoint, ListEndpoints] workflows: [ML Lifecycle Management] personas: [ML Engineer]