Red Hat blog
Organizations are using Artificial Intelligence and Machine Learning to improve the way they operate and deliver value to their customers. Red Hat OpenShift Data Science provides a model development environment delivered as a cloud service in combination with optional partner offerings to accelerate the time it takes to take a model from pilot to production.
Importance of AI/ML in organizations today
Organizations today across industries are harnessing the power of Artificial Intelligence (AI) and Machine Learning (ML) to drive innovation, improve their ability to compete, and increase customer and employee satisfaction. Access to the compute and storage capabilities required to gain insights from the massive amounts of data being collected has never been more available than it is today.
Core capabilities and components needed to get value out of AI/ML
Ambition and technology availability are two important parts of the equation but require an underlying platform which can provide secure, flexible, and scalable access to those resources to allow for rapid ideation and experimentation. Repeatability, transparency, and auditability are required to take experiments from pilot to production with confidence.
Flexibility (i.e., having options) and self-service capabilities are necessary in the rapidly changing world of AI/ML to provide data scientists access to innovation and advancements across the open source and competitive landscapes.
A long history of trusted platforms
Red Hat has a long history of providing trusted platforms on which organizations build and deploy business critical applications. Red Hat OpenShift puts developers and data scientists in control by providing an enterprise application platform powered by Kubernetes to build their AI/ML models, integrate them into applications, and deploy them at scale. And IT and operations are happy because of the security features present in Red Hat OpenShift, and the auditability and repeatability of deployments provided by the combination of OpenShift Pipelines, OpenShift GitOps, and infrastructure as code via Ansible.
How Red Hat OpenShift Data Science can help
Red Hat is excited to continue to build on this foundation of providing trusted solutions to enterprise grade problems with the Field Trial release of Red Hat OpenShift Data Science. OpenShift Data Science provides an AI/ML development and experimentation environment for data scientists as a cloud service on Red Hat OpenShift Dedicated and Red Hat OpenShift Service on AWS.
Data scientists can use the Jupyter notebook environment to build models using any of the provided notebook images with access to popular data science frameworks, including optimized images for Tensorflow and PyTorch. When the model is ready to deploy, data scientists can take advantage of the underlying OpenShift capabilities to deploy models as services via Source to Image (S2I).
ISV integrations and core capabilities
In addition to the core experimentation platform, Red Hat is helping organizations with end to end model operationalization through the integration of partner offerings to complement the core capabilities of the Red Hat OpenShift Data Science cloud service.
The basis of model development is data. For organizations looking for help accessing and integrating data sources across the hybrid cloud, Starburst Galaxy provides optimized data access across multiple data sources and platforms for use in model development.
Data scientists can extend the Juypter notebook environment and have access to the extensive set of data science frameworks and packages offered via Anaconda Enterprise. And for those that want access to the latest in AutoML, IBM Watson Studio provides an integrated suite to help build and manage models at scale.
For frameworks and deployments that can benefit from acceleration, whether deployed in the cloud or at the edge, Intel OpenVino provides a toolkit of pre-trained models optimized for Intel processors and GPUs. In the coming months, data scientists will have access to NVIDIA GPUs.
Once models are ready for deployment, users have the option to deploy and monitor those models via Seldon Deploy or IBM Watson Machine Learning.
These integrated partner offerings are available in addition to the full catalog of certified partners available via Red Hat Marketplace.
How to get started with OpenShift Data Science
In order to try OpenShift Data Science, you can get started quickly using the Red Hat Developer Sandbox. If you are interested in adding OpenShift Data Science to your OpenShift cloud service environment, please contact your Red Hat sales representative. Or just find out more about Red Hat OpenShift Data Science.
About the author
Steven Huels is a Software Development and Implementation Executive with a demonstrated track record leading multi-discipline organizations to achieve strategic objectives. Huels is known for building teams and growing market share through creativity and thought leadership in evaluating, setting direction, and successfully executing in response to market and organizational demands. Some areas of his expertise include Artificial Intelligence / Machine Learning, SaaS/PaaS/Big Data, and System Development/Integration.