Highlights:

  • At the AWS re:Invent event in Las Vegas, Nevada; Anyscale, the firm behind Ray open source – the unified computing platform for scaling any Machine Learning or Python workload, unveiled multiple advancements on the Anyscale PlatformTM.

Anyscale, the organization behind Ray open source, the unified computing framework for scaling any machine learning or Python workload, announced at the AWS re: Invent in Las Vegas, Nevada, many of their significant breakthroughs on the Anyscale PlatformTM.

The new features expand beyond the Ray open source’s advantages to make AI/ML and Python workload creation, experimentation and scalability simpler for the developers. Today, thousands of enterprises rely on Ray open source to scale AI workloads & applications and a rising number of these organizations are turning to the Anyscale PlatformTM for a streamlined Ray development and scaling experience.

The new Anyscale WorkspacesTM environment for rapid development and iteration is now available for early access. Workspaces offers a single and smoother developer experience for scaling machine learning (ML) workloads from a laptop to the cloud with no code modifications. Developers can now construct and deploy production-ready workloads in a single environment using familiar tools.

In addition, for accelerated development and rapid iteration, the Anyscale PlatformTM adds the ability to startup clusters up to 5x faster than the Ray open source, allowing developers to accelerate iteration, experimentation and deployments; job scheduling automation, including auto-scaling, alerting and more; and custom cluster environments to provide organizations with increased deployment and hosting flexibility.

Jake Carter, Director of Data, ML, and Technology at Biolexis Therapeutics said, “In the same time that it took to run our original workload – a week – we were able to effortlessly migrate all our Python workloads to the Anyscale Platform™, quickly fine-tune jobs for scaling, and move to production at scale effortlessly. It was remarkable and saved us a week end-to-end.”

Robert Nishihara, CEO, and Co-founder of Anyscale, said, “We are thrilled to see customers experience the benefits of the Anyscale Platform, which make Ray even more powerful and simple to use. Our customers have gained tremendous value from Anyscale, and I can confidently say that we’ve just touched the tip of the iceberg on making Ray even more impactful for developers and organizations who need to accelerate AI development and experimentation and remove the challenge of AI scaling.”

Howard Wright, VP and Global Head of Startups, AWS, said, “Enabling innovations like Anyscale Platform, which are igniting a golden age in AI and machine learning, is what the AWS tech stack was built for. Making it easier for companies to build machine learning models that are mature, reliable, and scalable with as little as two lines of code is the type of added value that we are excited to help bring to the market with Anyscale and Ray.”

New Capabilities and Highlights:

Workspaces delivers a single, laptop-like interface for developing and scaling ML workloads. Workspaces allow developers to continue utilizing familiar tools, such as VS Code, Jupyter, the terminal, and more while using the scalability and adaptability of the cloud. A developer can prepare data, adjust, train, and deploy workloads at any scale using a single script.

As one team in a manufacturing conglomerate said, “Anyscale Workspaces allows me to go from development to experimenting at scale, all the way to production, all within the same environment. Workspace reduces context switching for us by 50% and integrates easily with the other tools we use.”

Training and adjusting machine learning models is fundamentally iterative, and each iteration frequently needs cluster startup, tuning up, and shutdown. Anyscale shortens iteration cycles by reducing cluster setup events to less than two minutes, up to five times quicker than Ray open source.

Organizations may now deploy their unique Docker images as Anyscale cluster environments and use their current CI/CD workflows to develop and manage workloads running on Anyscale and Ray. This includes deploying Anyscale Workspaces, tasks and services using their own Docker tools and infrastructure.

Anyscale now offers a native method for scheduling jobs and connecting with leading orchestration solutions such as Airflow and Prefect. With task automation and integrations, Anyscale enables auto-scaling, alerting, and auto-retries, among other features, to facilitate and simplify the transfer of workloads to production.

Anyscale will exhibit at AWS re: Invent in Las Vegas, Nevada, from November 28 to December 2, 2022. Visit the Sands Expo Center at the Venetian Booth 1137 to discover more about Ray, a single compute framework for AI and Python scalability, and the Anyscale PlatformTM, a unified computing platform built on Ray. Discover why enterprises worldwide are embracing Anyscale and Ray to accelerate development and quickly scale AI/ML and Python workloads and apps.