Anyscale updates streamline cloud scaling for AI and ML developers

Distributed computing startup Anyscale Inc. announced today at AWS re:Invent that it is rolling out a series of updates to its platform aimed at enabling the development and scaling of workloads using Python-based artificial intelligence and machine learning for to simplify developers.

Anyscale is the company behind the open source Python framework Ray, used to run distributed computing projects. Ray includes both a general-purpose serverless programming interface for computer applications and an extended ecosystem of libraries. They enable developers to build scalable applications that can run on multicloud platforms without having to worry about the underlying infrastructure.

One of Ray’s key benefits is that it doesn’t require in-house distributed computing expertise.

The Anyscale cloud platform, on the other hand, is a managed version of Ray that aims to make it more accessible. The Ray platform requires a level of expertise that only a few high-level developers and computer specialists typically possess. Anyscale’s platform, running on AWS, solves the difficulty of taking an artificial intelligence prototype built on a laptop and scaling that model to hundreds of machines in the cloud.

Among the new features announced at re:Invent is the early access availability of the new Anyscale Workspaces environment, which aims to provide developers with a unified and more seamless experience when scaling machine learning workloads from a laptop to the cloud without the need for this is noteworthy code changes. Developers now have a single environment to build machine learning workloads and push them into production, Anyscale said.

One of the key benefits of Anyscale Workspaces is that developers can use the same familiar tools throughout the process, including VS Code and Jupyter, while reducing context switching when bringing new machine learning models to the cloud.

In a second update, the Anyscale platform gains the ability to launch clusters up to five times faster than the open-source Ray platform. As a result, developers can accelerate iteration, experimentation, and deployment of machine learning models, Anyscale said. Finally, Anyscale adds new automation features for job scheduling. With that, developers now have a native way to schedule jobs and integrate them with third-party orchestration tools like Airflow and Prefect, with auto-scaling, notification, auto-repeat, and other features available.

The updates are all about making machine learning developers faster and more productive, and early testers say they’re having the intended effect.

“In the same amount of time that we actually ran our original workload—one week—we were able to effortlessly migrate all of our Python workloads to the Anyscale platform, quickly fine-tune jobs to scale, and effortlessly go into production,” said he Jake Carter, director of data, machine learning and technology at Biolexis Therapeutics LLC. “It was remarkable and literally saved us a whole week.”

Howard Wright, AWS vice president and global head of startups, said that enabling innovations like the Anyscale platform is exactly what the AWS cloud was built to do. “Making it easy for companies to build mature, reliable, and scalable machine learning models with just two lines of code is the kind of value we’re excited to help bring to market with Anyscale and Ray,” he said.

Image: mamewmy/Freepik

Show your support for our mission by joining our community of experts, Cube Club and Cube Event. Join the community that includes Andy Jassy, ​​CEO of Amazon Web Services and, Michael Dell, Founder and CEO of Dell Technologies, Pat Gelsinger, CEO of Intel, and many more luminaries and experts.