Companies can use AI to build recommendation systems, help diagnose diseases and add self-driving capability to cars, according to Petuum co-founder, CEO and chief scientist Eric Xing, a computer science professor now on sabbatical from Carnegie Mellon University.
Now the start-up has raised $93 million in a new round of venture funding from SoftBank to develop and popularize this technology, Xing told CNBC.
Most organizations that want to use AI don't have access to a deep bench of technical talent. Nor do they have the computing infrastructure to run AI workloads at scale. AI involves training systems on massive amounts of data, which can stress general-purpose computers.
Petuum has created an operating system and software that make it easy to train and deploy artificial neural networks on the servers or other types of computers that they already use.
The software lets users choose from a variety of neural network architectures and then monitor the utilization of available computing resources. But the PetuumOS operating system and Poseidon AI framework are still in beta testing.
Many venture-backed AI software companies are developing systems that serve one niche, like computer vision for self-driving cars, or ad-buying algorithms for agencies and brands.
Petuum is taking a "horizontal" approach, Xing said. Its technologies are meant to be applied to any industry. Google and Microsoft also offer cloud services that can be used for a wide variety of AI applications.
Petuum sometimes serves as a consultancy that develops and maintains new AI and machine learning solutions for other businesses. But it also plans to sell or license software that can be used on a "self-serve" basis.