British cloud service company Civo has announced “Kubeflow as a Service.” This project allows companies to use machine learning technology on a large scale without having to buy hardware themselves.
Kubeflow as a Service is the first product in the line of the new Civo ML platform, which operates on an open-source basis. The company has spotted a gap in the market when it comes to accessing large-scale GPU infrastructure. Few organizations have the capital to invest in a cluster of GPUs capable of machine learning effectively. With Civo’s model, smaller companies merely rent the hardware needed for ML purposes when needed. Civo itself says it is fully supported by Kubernetes.
On the shoulders of giants
Civo says collaboration is a high priority with the new initiative. The open-source nature of the project makes it easy for developers to make changes for specific ML purposes. This allows small groups of developers to “stand on the shoulders of giants”, as the company puts it. Indeed, Kubeflow as a Service is designed to take a lot of preparatory work away from ML developers.
Chief Innovation Officer at Civo Josh Mesout emphasizes the importance of open source in this project. “Evaluation is integral to continued success, and we’re keen to shape the platform around the users’ feedback and grow the service to answer developers’ needs.”
Also read: Events, dollars & cents: the state of open source in 2023