---
id: azureml-pipeline-steps
version: "1.62.0"
license: https://aka.ms/azureml-sdk-license
license_treatment: unclear
maintenance: active
---

# azureml-pipeline-steps — Aeva : represents a unit of computation in azureml-pipeline

License: unclear · Maintenance: active · Downloads: 230.2K/mo

## What it is and what it does

azureml-pipeline-steps is a component of the Azure Machine Learning SDK that defines the step types used to construct pipelines—units of computation that can be orchestrated and executed on Azure ML infrastructure. It sits between azureml-pipeline-core (which manages the pipeline graph) and azureml-train-core and azureml-train-automl-client (which provide the training and AutoML execution logic). The package is production-stable, supports Python 3.8–3.11, and has been actively maintained since its 2018 release.

You install it when you need to programmatically define and run machine learning workflows on Azure—for example, to chain data preparation, model training, and evaluation steps into a reproducible pipeline. It is not a standalone tool; it is part of the larger azureml ecosystem and requires an Azure ML workspace to execute pipelines.

Use it for:

- Define a multi-step training pipeline that preprocesses data, trains a model, and evaluates it on Azure ML.
- Orchestrate automated machine learning (AutoML) jobs as steps within a larger pipeline workflow.
- Build reproducible ML workflows that can be scheduled, versioned, and shared across a team.
- Chain custom Python scripts and pre-built training components into a single executable pipeline.
- Integrate data processing steps with model training in a single Azure ML pipeline definition.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Defines step types for building computation units in Azure Machine Learning pipelines, enabling orchestration of training, data processing, and automated ML workflows.

Yes, if you are building machine learning pipelines on Azure ML. It is a required component of the Azure ML SDK for pipeline orchestration, has low install friction, is actively maintained, and has no known vulnerabilities. The unclear license requires review before use in proprietary contexts, but it is standard for Microsoft Azure SDKs. Install it as part of your Azure ML workflow, not as a standalone tool.

## Install

pip install azureml-pipeline-steps
uv add azureml-pipeline-steps
poetry add azureml-pipeline-steps

## Installing azureml-pipeline-steps

Before you install: Low install friction with a pure Python wheel. Maintenance status is active with a recent release (170 days ago). Depends on three azureml packages (azureml-train-core, azureml-train-automl-client, azureml-pipeline-core), which may add transitive dependencies to your environment.

License in practice: License treatment is unclear—the package points to a Microsoft-hosted license URL rather than a standard SPDX identifier. Review the linked license before use in proprietary or restricted contexts.

Quickstart:

pip install azureml-pipeline-steps

from azureml.pipeline.steps import PythonScriptStep
from azureml.pipeline.core import Pipeline

# Define a step and add it to a pipeline
step = PythonScriptStep(script_name='train.py')
pipeline = Pipeline(steps=[step])

Requires Python 3.8 or later (supports 3.8, 3.9, 3.10, 3.11). Intended for use within an Azure ML workspace or SDK environment.

Verify before relying:

- Whether the Microsoft license URL resolves to a standard open-source or proprietary license agreement.
- Whether azureml-train-core and azureml-train-automl-client introduce heavy or optional transitive dependencies.
- Specific step types and APIs available in version 1.62.0 beyond the generic 'unit of computation' description.

## Package facts

- License: https://aka.ms/azureml-sdk-license (unclear)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 230.2K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

azure ml pipeline steps, azureml pipeline computation units, azure machine learning workflow steps, ml pipeline orchestration, azure automl pipeline, machine learning pipeline components, azure-ml, pipeline-orchestration, ml-workflow

[View on SkillFed](https://skillfed.io/packages/azureml-pipeline-steps) · [View on PyPI](https://pypi.org/project/azureml-pipeline-steps/)