Kubeflow

A major contributor to this article appears to have a close connection with its subject. It may require cleanup to comply with Wikipedia's content policies, particularly neutral point of view. Please discuss further on the talk page. (September 2022) (Learn how and when to remove this message)

Kubeflow

Original author(s)	Google
Developer(s)	Kubeflow Contributors[1] - AWS, Bloomberg, Google, IBM, NVIDIA, Nutanix, Red Hat, Arrikto, and others
Initial release	April 5, 2018; 7 years ago (2018-04-05)[2]
Stable release	1.10[3] / April 1, 2025; 37 days ago (2025-04-01)
Repository	github.com/kubeflow
Written in	Go, Python, TypeScript
Platform	Kubernetes
Type	Machine Learning Platform
License	Apache License 2.0
Website	kubeflow.org

Kubeflow is an open-source platform for machine learning and MLOps on Kubernetes introduced by Google. The different stages in a typical machine learning lifecycle are represented with different software components in Kubeflow, including model development (Kubeflow Notebooks^[4]), model training (Kubeflow Pipelines,^[5] Kubeflow Training Operator^[6]), model serving (KServe^[a][7]), and automated machine learning (Katib^[8]).

Each component of Kubeflow can be deployed separately, and it is not a requirement to deploy every component.^[9]

The Kubeflow project was first announced at KubeCon + CloudNativeCon North America 2017 by Google engineers David Aronchick, Jeremy Lewi, and Vishnu Kannan^[10] to address a perceived lack of flexible options for building production-ready machine learning systems.^[11] The project has also stated it began as a way for Google to open-source how they ran TensorFlow internally.^[12]

The first release of Kubeflow (Kubeflow 0.1) was announced at KubeCon + CloudNativeCon Europe 2018.^[13][14] Kubeflow 1.0 was released in March 2020 via a public blog post announcing that many Kubeflow components were graduating to a "stable status", indicating they were now ready for production usage.^[15]

In October 2022, Google announced that the Kubeflow project had applied to join the Cloud Native Computing Foundation.^[16][17] In July 2023, the foundation voted to accept Kubeflow as an incubating stage project.^[18][19]

Machine learning models are developed in the notebooks component called Kubeflow Notebooks. The component runs web-based development environments inside a Kubernetes cluster, with native support for Jupyter Notebook, Visual Studio Code, and RStudio.^[20]

Once developed, models are trained in the Kubeflow Pipelines component. The component acts as a platform for building and deploying portable, scalable machine learning workflows based on Docker containers.^[21] Google Cloud Platform has adopted the Kubeflow Pipelines DSL within its Vertex AI Pipelines product.^[22]

For certain machine learning models and libraries, the Kubeflow Training Operator component provides Kubernetes custom resources support. The component runs distributed or non-distributed TensorFlow, PyTorch, Apache MXNet, XGBoost, and MPI training jobs on Kubernetes.^[6]

The KServe component (previously named KFServing^[23]) provides Kubernetes custom resources for serving machine learning models on arbitrary frameworks including TensorFlow, XGBoost, scikit-learn, PyTorch, and ONNX.^[24] KServe was developed collaboratively by Google, IBM, Bloomberg, NVIDIA, and Seldon.^[23] Publicly disclosed adopters of KServe include Bloomberg,^[25] Gojek,^[26] the Wikimedia Foundation,^[27] and others.^[28]

Lastly, Kubeflow includes a component for automated training and development of machine learning models, the Katib component. It is described as a Kubernetes-native project and features hyperparameter tuning, early stopping, and neural architecture search.^[29]

• Official website
• Kubeflow on GitHub