Google Announces Kubeflow to Bring Kubernetes to Machine Learning
The fully open source project is designed to help engineers build a machine learning stack using Kubernetes.
After Kubernetes and TensorFlow, Google has now released Kubeflow, a new open source project that makes it easy to consume machine learning (ML) stacks with Kubernetes.
Kubernetes is being touted as the cloud Linux, and an increasing number of people are employing it in different use cases. Machine learning is one of the fastest growing use cases for Kubernetes, but it's quite a challenge to get the entire machine learning stack up and running.
“Building any production-ready machine learning system involves various components, often mixing vendors and hand-rolled solutions. Connecting and managing these services for even moderately sophisticated setups introduces huge barriers of complexity in adopting machine learning,” said David Aronchick and Jeremy Lewi, Project Manager and Engineer, respectively, on the Kubeflow project. “Infrastructure engineers will often spend a significant amount of time manually tweaking deployments and hand rolling solutions before a single model can be tested.”
Kubeflow solves this problem because it makes using ML stacks on Kubernetes fast and extensible. It’s hosted on GitHub, and the repository contains three components: JupyterHub, to create and manage interactive Jupyter notebooks; a TensorFlow (TF) Custom Resource Definition (CRD) that can be configured to use CPUs or GPUs and adjusted to the size of a cluster with a single setting; and a TF Serving container.
Kubeflow is a muticloud solution, and if you can run Kubernetes in your environment, you can run Kubeflow.
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Subscribe to our ADMIN Newsletters
Support Our Work
Linux Magazine content is made possible with support from readers like you. Please consider contributing when you’ve found an article to be beneficial.

News
-
There's a New Linux AI Assistant in Town
Newelle is a Linux AI assistant that can work with different LLMs and includes document parsing and profiles.
-
Linux Kernel 6.16 Released with Minor Fixes
The latest Linux kernel doesn't really include any big-ticket features, just a lot of lines of code.
-
EU Sovereign Tech Fund Gains Traction
OpenForum Europe recently released a report regarding a sovereign tech fund with backing from several significant entities.
-
FreeBSD Promises a Full Desktop Installer
FreeBSD has lacked an option to include a full desktop environment during installation.
-
Linux Hits an Important Milestone
If you pay attention to the news in the Linux-sphere, you've probably heard that the open source operating system recently crashed through a ceiling no one thought possible.
-
Plasma Bigscreen Returns
A developer discovered that the Plasma Bigscreen feature had been sitting untouched, so he decided to do something about it.
-
CachyOS Now Lets Users Choose Their Shell
Imagine getting the opportunity to select which shell you want during the installation of your favorite Linux distribution. That's now a thing.
-
Wayland 1.24 Released with Fixes and New Features
Wayland continues to move forward, while X11 slowly vanishes into the shadows, and the latest release includes plenty of improvements.
-
Bugs Found in sudo
Two critical flaws allow users to gain access to root privileges.
-
Fedora Continues 32-Bit Support
In a move that should come as a relief to some portions of the Linux community, Fedora will continue supporting 32-bit architecture.