Key features

Structure pipelines based on components

  • Create code modules that can be reused across projects
  • Cache pipeline runs to avoid repetition of data processing
  • Warm-start models

Easy-to-use interface to generate models

  • Use tf/keras code or select one of the templated model definitions
  • Configure your model and the training process according to your specifications

Repeatable and transparent pipelines

  • Trace easily what pipeline configurations lead to which results
  • Version pipeline runs through git commits for repeatability
  • Conduct hyperparameter tuning and compare experiments

Comprehensive evaluation of your models

Get early access

Be the first to see the platform in action.

Proudly powered by: