Anyscale, a startup just launched last year that aims to make distributed computing dead-simple, just raised a new round of funding to build out its sales and marketing teams.
A team of Berkeley researchers founded the firm around Ray, an open source project they created that makes it easy for developers to distribute compute and memory across multiple computers. This is particularly important for any application that uses machine learning, which requires immense amounts of power and memory.
"If you ask every developer, this is like the holy grail," cofounder and executive chairman Ion Stoica told Business Insider.
Anyscale raised a $40 million Series B funding from NEA, with participation from Andreessen Horowitz, Intel Capital, and Foundation Capital, the firm announced Wednesday. It has now raised $60.6 million in total, though declined to share its valuation, and plans to use the new funding to keep up with growing demand.
Anyscale's system gives developers "the same convenience and same familiarity as using a laptop" as they build and run their applications, according to Stoica.
"Building these applications is incredibly hard," he said. "That's why Uber and Google and Microsoft have an army of engineers who are super-experts building these applications. Our goal is to make this easy, to make it possible for programmers to build these applications."
The Ray open source project is growing in popularity
Ray's userbase has been growing quickly, even during the pandemic, according to cofounder and CEO Robert Nishihara. For example, more developers are building libraries of code to support Ray and make it easier to train machine learning algorithms.
"The high level trend I'm betting on is: Distributed computing is becoming the norm," Nishihara said. "It's becoming the default way of doing things. You want to build these scalable applications, you're hiring an infrastructure team. Not very many developers learn how to do that. At the same time, tons of people are learning to program in Python, tons of people are learning to program in general."
Burger King has used Ray to build apps that give better recommendations to customers ordering food. Likewise, Amazon Web Service's AI platform SageMaker uses Ray for helping users get started with reinforcement learning, a branch of AI that trains algorithms by "rewarding" them for performing well and giving penalties for performing incorrectly.
"I just want to reinforce the fact that we really are seeing many customers adopt machine learning at scale," Bratin Saha, vice president of machine learning at AWS, told Business Insider. "This collaboration with Ray is just one part of making SageMaker the place to do machine learning in the cloud."
Anyscale plans to use the new funding to scale its product to new potential customers, continue investing in open source, and building its team with both engineering and non-engineering roles.
"We are just at the beginning," Stoica said. "We are going to have a lot of investment and engineering to achieve [our] vision."
Got a tip? Contact this reporter via email at rmchan@jkmperu.com, Signal at 646.376.6106, Telegram at @rosaliechan, or Twitter DM at @rosaliechan17. (PR pitches by email only, please.) Other types of secure messaging available upon request.
