Repolode
Submit
databricks

databricks/mlops-stacks

A Databricks asset bundle template that scaffolds ML projects with production-ready code, resource configs, and CI/CD pipelines.

View on GitHubWebsite
Stars
725
Forks
268
Open issues
16
Last commit
19 days ago

What is it?

MLOps Stacks is a project template for Databricks that generates a new machine learning project with the structure, configuration, and deployment automation needed to move models from development to production. It is distributed as a Databricks CLI bundle template, so a single `databricks bundle init mlops-stacks` command creates the project from user-supplied parameters.

The generated project contains three modular parts: example ML code (training and batch inference modules with unit tests), ML resources defined as code through Databricks CLI bundles so that job and pipeline changes go through pull requests, and CI/CD workflows for GitHub Actions or Azure DevOps that test code and deploy it across dev, staging, and prod workspaces. Teams can set up the project alone, set up CI/CD alone, or copy individual components into existing projects.

The stack is customizable and works for Databricks resources beyond ML, such as Delta Live Tables pipelines. It is currently in public preview and requires Python 3.8+ and Databricks CLI v0.236.0 or later.

Key features

  • 01Generates a new ML project from a Databricks bundle template with one command.
  • 02Includes example training and batch inference code with unit-tested Python modules.
  • 03Defines ML jobs and resources as code via Databricks CLI bundles, changed through pull requests.
  • 04Provides CI/CD workflows for GitHub Actions, GitHub Enterprise Server, Azure DevOps, and GitLab.
  • 05Supports separate dev, staging, and prod workspaces, and Unity Catalog model registration.
  • 06Offers an optional Feature Store component.

Use cases

  • Starting a new ML project on Databricks with CI/CD already configured.
  • Adding production deployment automation to an existing ML project by copying its resource and workflow directories.
  • Managing Databricks jobs and pipelines as version-controlled code across environments.

Quick start

  1. 1

    Install the Databricks CLI (version 0.236.0 or later)

  2. 2

    Create a new project from the MLOps Stacks template

    databricks bundle init mlops-stacks

Things to know

  • Requires Python 3.8+ and Databricks CLI 0.236.0 or later.
  • The feature is in Databricks public preview.
  • Requires Databricks workspaces to generate and run the ML pipelines.

In collections

Testing frameworks & test automation tools

Topics

#mlops#machine-learning#databricks#ci-cd#infrastructure-as-code#project-template

FAQ

What is mlops-stacks?
A Databricks asset bundle template that scaffolds ML projects with production-ready code, resource configs, and CI/CD pipelines. MLOps Stacks is a project template for Databricks that generates a new machine learning project with the structure, configuration, and deployment automation needed to move models from development to production. It is distributed as a Databricks CLI bundle template, so a single `databricks bundle init mlops-stacks` command creates the project from user-supplied parameters.
Is mlops-stacks free to use? What license does it have?
mlops-stacks is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is mlops-stacks still maintained?
Actively maintained. The last commit to mlops-stacks was on Oct 20, 2026.
What can I use mlops-stacks for?
Starting a new ML project on Databricks with CI/CD already configured. Adding production deployment automation to an existing ML project by copying its resource and workflow directories. Managing Databricks jobs and pipelines as version-controlled code across environments.
How do I install mlops-stacks?
Install the Databricks CLI (version 0.236.0 or later) Create a new project from the MLOps Stacks template: databricks bundle init mlops-stacks
Where can I get mlops-stacks?
The source code and installation instructions are on GitHub: https://github.com/databricks/mlops-stacks

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects