Red Hat launches a managed OpenShift Data Science service to enable organizations in using machine learning capabilities.
With the move, Red Hat offers data scientists a customizable machine learning platform. Red Hat’s solution includes a collection of open-source tools and technology that allows the development of experimental models. In doing so, data scientists should not have to worry about underlying infrastructure or lock-in from a specific cloud provider.
The platform is extensible with capacity and machine learning models from various Red Hat partners through the Red Hat Marketplace. Developed models can be exported in a format suitable for container environments, making them suitable for hybrid and edge cloud environments.
The managed service features AI, machine learning development and visualization tools such as Jupyter Hub, TensorFlow and PyTorch. Optional tools include Anaconda Commercial Edition and IBM Watson Studio. Starburst is supported for data engineering activities. Red Hat OpenShift Streams and Apache Kafka suit data streaming. Modelling is done with the OpenShift Source-to-Image tool, Red Hat OpenShift API Management or Seldon Deploy. Hardware acceleration takes place via Nvidia with GPU Operator, the Intel OpenVINO toolkit or, optionally, the Intel oneAPI AI Analytics Toolkit.
Red Hat OpenShift Data Science is now available as a test add-on for Red Hat OpenShift Dedicated and the Red Hat OpenShift Service for AWS. Users get complete support and only pay the cost of underlying infrastructure.