From left: Petuum co-founders Eric Xing, Qirong Ho, and Ning Li

Petuum in CB Insights’ 2018 AI 100

We couldn’t be more delighted to be named to the CB Insights AI 100 list for the second time!

Our mission is unique among Artificial Intelligence startups and we’re incredibly thankful that our solution is resonating with industry experts. At Petuum, we are working to democratize the processes underlying the creation of AI systems and make even the most advanced AI technology accessible and affordable for everyone who needs it. Our end goal is to dramatically lower the barrier to AI adoption and allow for the integration of practical AI into every industry.

Most companies break into the AI industry with an amazing idea, like a car that drives itself or a computer that recognizes faces. The product or function they create is beautiful and novel; however, it is expensive to produce, has a niche function, and is often not repeatable or suitable for mass production. Such AI products are exciting and revolutionary, but their benefits are limited.

By virtualizing and standardizing AI and ML programming and computing, we want to transform it into a simple-to-use, transparent, repeatable, and everyday tool for all enterprises.

Currently, AI development is exclusive to tech industry giants — high costs and limited talent are major barriers to widespread adoption. The pool of talent with sufficient training, experience, and expertise in this field is small, largely due to the relative newness of the work and the level of training necessary to understand AI, let alone create it. In such a small talent pool, every capable expert is most likely to be recruited to one of the giant companies that lead the AI market since they 1) can pay more, 2) guarantee cutting-edge work, and 3) have name-brand prestige and publicized progress.

Luckily, like AI itself,* Petuum was born at CMU,** bypassing Silicon Valley’s AI talent shortage. Our team of expert CMU alums banded together to address AI’s inaccessibility by building a platform of standard AI building blocks that can be further customized by the companies that use them. This will eventually allow for the integration of AI into the basic infrastructure of all industries — not just the tech industry, and not just large, public companies.

Accessible and affordable AI will drive higher productivity, better service, lower costs, and faster delivery across all applications.

With pick-and-choose, customizable AI, both enterprise and small businesses will be able to achieve the specialized products they envision. With ready-to-use, no-experience-needed AI, the technology will be accessible to those outside of the tech industry with little-to-no experience with programming. Traditional industries like retail, healthcare, and finance can benefit from faster processing, smarter programs, and improved operations.

Thank you, CB Insights, for recognizing the importance of our mission and supporting us!


*On Thursday, December 15, 1955, Allen Newell and Herbert Simon, researchers at Carnegie Institute of Technology (now Carnegie Mellon University), created the first artificial intelligence program, the “Logic Theorist.” They called it a “thinking machine,” describing a computer that solved problems heuristically in remarkable likeness to the way humans solve problems.

**Carnegie Mellon University (CMU) has been widely recognized as the center of many breakthroughs in AI and a prime source of new talent. Therefore, unsurprisingly, CMU researchers have been subject to aggressive recruiting by tech industry giants — Delphi Automotive PLC acquired the university’s spinoff, Ottomatika; Uber lured graduates and staff to work on autonomous vehicles with its Advanced Technologies Group; and Google (temporarily) attained a top mind for fraud prevention and ad targeting research.