Cnvrg.io develops on-demand service to run AI workloads across infrastructures

Cnvrg.io, the Intel-owned company that offers a platform to help data scientists build and deploy machine learning applications, has opened early access to a new managed service called Cnvrg.io Metacloud.

The offering, as the company explains, gives AI developers the flexibility to run, test, and deploy AI and machine learning (ML) workloads on a mix of mainstream infrastructure and hardware choices, even within the same AI/ML workflow or pipeline.

Cnvrg.io Metacloud: Flexibility for AI developers

AI experts can often find themselves struggling to scale their projects due to the limitations of the cloud or on-premise infrastructure in use. They do get the option to switch to a new environment, but that means re-instrumenting a completely new stack as well as spending a lot of cash and time. This eventually keeps most of the users locked on a single vendor, making it a major obstacle to scaling and operationalizing AI.

Cnvrg.io Metacloud tackles this challenge with a flexible software-as-a-service (SaaS) interface, where developers can pick cloud or on-premise compute resources and storage services of their choice to match the demand of their AI/ML workloads.

The solution has been designed using cloud-native technologies such as containers and Kubernetes, which enables developers to pick any infrastructure provider from a partner menu to run their project. All users need to do is create an account, select the AI/ML infrastructure (any public cloud, on-premise, co-located, dev cloud, pre-release hardware, and more), and run the workload, the company said.

Plus, since there is no commercial commitment, developers can always change to a different infrastructure to meet growing project demands or budget constraints. The current list of supported providers includes Intel, AWS, Azure, GCP, Dell, Redhat, VMWare, and Seagate.

“AI has yet to meet its ultimate potential by overcoming all the operational complexities. The future of machine learning is dependent on the ability to deliver models seamlessly using the best infrastructure available,” Yochay Ettun, CEO and cofounder of Cnvrg.io, said in a statement.

“Cnvrg.io Metacloud is built to give flexibility and choice to AI developers to enable successful development of AI instead of limiting them, so enterprises can realize the full benefits of machine learning sooner,” he added.

Cnvrg.io Metacloud will be provided as part of the Cnvrg.io full-stack machine learning operating system, designed to help developers build and deploy machine learning models. The early access version of the solution can be accessed upon request via the company website.

Notably, this is the first major announcement from Cnvrg.io since its acquisition by Intel in 2020. Prior to the deal, the company had raised about $8 million from multiple investors, including Hanaco Venture Capital and Jerusalem Venture Partners.