DEVOPSdigest

Red Hat announced new capabilities for Red Hat OpenShift AI.

Building and expanding upon the proven capabilities of Red Hat OpenShift and Red Hat OpenShift Data Science, Red Hat OpenShift AI provides a consistent, scalable foundation based on open source technology for IT operations leaders while bringing a specialized partner ecosystem to data scientists and developers to capture innovation in AI. To that end, Red Hat OpenShift AI underpins the generative AI services of IBM watsonx.ai, IBM’s artificial intelligence platform designed to scale intelligent applications and services across all aspects of the enterprise, fueling the next generation of foundation models.

OpenShift AI provides the infrastructure consistency across training, deployment and inference to unlock the potential of AI.

Red Hat OpenShift AI provides a standardized foundation for creating production AI/ML models, as well as running the resulting applications. OpenShift AI delivers the consistency, ease-of-use and cloud-to-edge deployment options of Red Hat OpenShift. Platform engineers can create scalable configurations, specific to the needs of their data scientists and developers. OpenShift AI can be used to fine-tune existing models, saving on initial model training costs, increasing time to value and expanding AI usage beyond specialized data scientists.

Red Hat OpenShift AI provides several optional technology partner offerings including Anaconda, IBM Watson Studio, Intel OpenVINO and AI Analytics Toolkit, NVIDIA AI Enterprise and Starburst. It also provides access to 30 additional AI/ML certified partners as part of the OpenShift ecosystem.

Customers with specific regulatory and compliance requirements, including air-gapped and disconnected environments, can prepare data, and develop, train and deploy models on-premises using OpenShift AI. Customers can also develop models in the public cloud, and deploy them on-premises or at the edge using the same consistent tooling and interfaces, providing a unique, hybrid MLOps environment that enables collaboration between IT operations, data science and application developers.

The proven foundation of OpenShift AI enables customers to more reliably scale to train foundation models using OpenShift’s native GPU acceleration features on-premises or via a cloud service.

Recent enhancements to Red Hat OpenShift AI include:

- Deployment pipelines for AI/ML experiment tracking and automated ML workflows, which helps data scientists and intelligent application developers to more quickly iterate on machine learning projects and build automation into application deployment and updates.

- Model serving now includes GPU support for inference, and custom model serving runtimes that improve inference performance and improved deployment of foundation models.

- Model monitoring enables organizations to manage performance and operational metrics from a centralized dashboard.

Chris Wright, CTO and SVP, Global Engineering, Red Hat, said: “Foundation models provide real, tangible benefits to enterprises when it comes to harnessing the benefits of AI, but they still require investment in training and fine-tuning to meet the unique needs of an enterprise. Red Hat’s vision for enterprise AI builds on this existing reality with Red Hat OpenShift AI, which provides a flexible and scalable foundation to train, maintain, fine-tune and actually use foundation models in production. Best of all, OpenShift AI is still OpenShift, meaning that IT organizations trust it and understand it, and can extend their AI/ML operations from meeting today’s needs to tomorrow’s.”