DKube provides the tools for groups of users to collaborate on projects in a variety of ways. Multi-tenancy is native to the platform, allowing the users within an organization to securely share code, datasets, models, images, pipelines, and resources.

The collaboration features of DKube extend what is available from the standard Kubeflow framework, and are integrated seamlessly into the UI-based workflow. A user can take work from another group member and use it for enhanced analysis or training. A user can even compare the metrics from models that were created by different users through the powerful display and compare feature based on MLFlow.

Within the data science workflow, users can optionally partition entities such as code, datasets, & models, etc into projects. This provides an intuitive way to show only the entities of interest on the display.

DKube also allows groups of users to cooperate closely in order to solve a specific problem. The owner of the problem provides the training and test datasets, and identifies the metric goals. Multiple users independently create and submit solutions to the problem. The solutions can be viewed on a dashboard in order to understand which solutions best address the goals.