Klever is a cloud native platform for machine learning workloads. It help users to train, distribute, manage and serve their machine learning models. Klever is built on top of many state-of-art open source projects like Kubeflow, Harbor and so on. Its current components are:

• ormb: A CLI and a library to help users manage machine learning models like Docker images.
• Model registry: A centralized registry, set of APIs and UI ( Will be added soon ) for machine learning models, based on Harbor and ormb.
• Serving platform: A platform to deploy models in the model registry with the help of Seldon Core.

Components to be added:

• Training platform.

Architecture

Roadmap

Please have a look at Roadmap.