sagemaker-mlops

What it is and what it does

sagemaker-mlops is the orchestration layer of the SageMaker SDK, sitting above the Core, Train, and Serve packages. It provides pipeline definitions, step implementations (AutoML, model creation, bias checks, EMR, Lambda, batch transform monitoring, and others), and configuration utilities for parallel execution, retry policies, and selective execution. The package resolves architectural constraints by centralizing workflow logic that needs to import from multiple lower-level SageMaker components.

You use it to define and execute multi-step machine learning workflows on Amazon SageMaker. It exposes classes for pipelines, steps, and specialized step types, along with configuration objects for controlling execution behavior. The package also re-exports ModelBuilder from sagemaker-serve for convenience. It requires Python 3.10 or later and depends on boto3, botocore, cryptography, pyiceberg, pyarrow, and s3fs alongside the three core SageMaker packages.

Use it for:

• Define multi-step ML pipelines with conditional logic, parallel execution, and retry policies for SageMaker workflows
• Orchestrate training, model evaluation, bias checking, and deployment steps in a single declarative pipeline
• Integrate Lambda functions, EMR clusters, and batch transform jobs into SageMaker workflow definitions
• Configure selective execution and parallelism settings for large-scale model training and serving workflows
• Build and register models with quality and bias checks as part of an automated MLOps pipeline

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides high-level orchestration for Amazon SageMaker workflows, including pipeline definitions, step implementations, and model building utilities that coordinate training, serving, and core SageMaker components.

Yes, if you are building machine learning workflows on Amazon SageMaker. The package is actively maintained, has low install friction, carries a permissive Apache 2.0 license, and integrates cleanly with the modular SageMaker SDK architecture. The Alpha status suggests it is still evolving; verify stability requirements for your use case before production deployment.

Install

sagemaker-mlops on PyPI

pip install sagemaker-mlops

uv add sagemaker-mlops

poetry add sagemaker-mlops

Installing sagemaker-mlops

Before you install

Low install friction; pure Python wheel with no compiled dependencies. Actively maintained as of 3 days ago. Depends on 9 runtime packages including sagemaker-core, sagemaker-train, and sagemaker-serve, which form the SageMaker SDK's modular architecture.

License in practice

Apache License 2.0 (permissive). You may use, modify, and distribute this package freely in commercial and open-source projects, provided you include the license text and note any modifications you make.

Quickstart

pip install sagemaker-mlops

from sagemaker.mlops import Pipeline

pipeline = Pipeline(name="my-pipeline", steps=[])
pipeline.upsert()
pipeline.start()

Requires Python 3.10 or later. Requires AWS credentials and SageMaker permissions to execute pipelines. The 9 runtime dependencies (sagemaker-core, sagemaker-train, sagemaker-serve, boto3, botocore, cryptography, pyiceberg, pyarrow, s3fs) must be available.

Verify before relying

• Whether all 9 runtime dependencies are automatically resolved or must be installed separately
• Whether this package is intended for development/editable install or production use
• Stability guarantees given the 'Alpha' development status classifier
• Concrete usage examples for specific step types beyond what the fact sheet documents

Package facts

Evidence: sagemaker_mlops-1.19.0-py3-none-any.whl

Tags

More Artificial Intelligence packages

Further reading