An AI Idea to Production Cycle: Cyxtera Works With NVIDIA to Deliver Speed and Accuracy Where It Counts

David Murphy • November 15, 2021 • 4 minute read

AI/ML, Enterprise Bare Metal, Channel Partners in Business

Artificial intelligence represents a major growth area for startups. According to research, investors have invested $29.5 billion into AI startups in the first half of this year, surpassing the total for 2020, which was $27.8 billion. The investment community is on pace to crush the level of support, year over year.

With so much attention and resources focused on AI solutions, the expectations for their impact are incredibly high. However, there are challenges that all AI startups face in scaling up their innovations.

In the AI/ML segment, scaling is far more difficult due to the need for powerful computer processing that can rapidly process AI calculations in parallel at varying degrees of precision. As AI models grow in scope and complexity, data points can easily run into the thousands. In a pilot stage, a company may release significant breakthroughs delivered for a limited dataset. However, the true measure of success requires solutions to meet the demands of processing data that may be in the hundreds of thousands, or even larger, and when running AI inference, the data is often processed in real-time.

As shared in a recent VentureBeat article, “[Our] research has found that only half of AI projects make it from pilot into production, and those that do take an average of nine months to do so,” said Svetlana Sicular, research VP at Gartner. “Innovations such as AI orchestration and automation platforms and model operationalization are enabling reusability, scalability, and governance, accelerating AI adoption and growth.”

More than ever, speed to market is an essential requirement in today’s business world. Innovators are creating startup companies at an aggressive pace to develop new technologies. For startups focused on AI solutions, finding the best approaches to scale up are essential. This business dynamic has created opportunities for Shakudo.

Shakudo offers Hyperplane, an end-to-end platform that helps teams drive AI ideas to production without having to rely on extensive engineering support. Hyperplane fits into a company’s existing architecture and infrastructure with relative ease. Data scientists scale models from small data to terabytes of data using one line of code. No more Docker files or configurations to get models to production. The team at Shakudo created Hyperplane so that data scientists may produce and maintain their own models through to the production cycle easier and faster than before. As a result, data scientists may scale up models from development data to production data, including iterating models on real data.

Built for emerging tech, Shakudo is focused on providing essential infrastructure support for technologies that are still in early adoption stages across the industry, with a sharp focus on reducing friction and providing a great user experience.

In order to meet these objectives, the leadership team at Shakudo needed more accelerated GPU computing power, but in a managed and controlled manner. The appeal of the Cyxtera AI/ML compute as a service offering featuring NVIDIA DGX systems was clear as it enables Shakudo and its clients to access on-demand AI compute without time-consuming procurement and configuration steps. The clients’ business needs required them to move fast, and they needed access to resources that would accelerate their time to market. Financially, the solution made great sense because the on-demand subscription service can scale up and down at the pace the Shakudo business requires.

As a cutting-edge startup, Shakudo is a proud member of NVIDIA Inception, a free program designed to help startups evolve faster through access to technical training, connections with venture capitalists, and co-marketing support. Moving forward, Shakudo is confident they have the IT solutions to help them increase the number of converted pilot projects and clients. Collaborating with Cyxtera and NVIDIA Inception, Shakudo has been able to fine-tune their product roadmap, as well as overall marketability.

Speed to Solution

Shakudo joined NVIDIA Inception in March of 2021 and was eager to move forward with a cost- and time-effective way to bring machine learning tools beyond PoC to production for validation. In just over six months, Shakudo will have its program activated.

Hyperplane makes it easier to deploy, scale and maintain AI software. With built-in packages, tools and optimizations, scaling AI solutions is as easy as a single line of code with the OpEx solution. When speed counts – and, these days, it always does – the team at Shakudo developed the solutions needed to reduce costs, shorten timelines with PoCs, and offer experimental code that scales without rework.

Hyperplane is an end-to-end environment where scientists can experiment, test, deploy, monitor and troubleshoot their own work. Hyperplane exposes robust and consumable DevOps-friendly GraphQL APIs so teams can easily interact with AI solutions deployed on Hyperplane.

The result? An improved developer experience for both data scientists building AI solutions and developers that use the solutions to build products.

Startups can apply for free to join NVIDIA Inception, no matter what their current funding stage is. There are no application deadlines, cohorts, or term limits. To learn more about Cyxtera’s support of the NVIDIA Inception program, please visit here.

Views and opinions expressed in our blog posts are those of the employees who made them and do not necessarily reflect the views of the Company. A reader should not unduly rely on any statements made therein.

David Murphy

Director, Channel & Partner Marketing