Deloitte Makes Big Nvidia CUDA AI Play With OpTeamizer Acquisition

‘They work really closely with Nvidia’s CUDA stack. They can help optimize AI models for running on Nvidia GPUs, where, in the market today, a lot of AI is running. Really, what they're doing is these high-performance optimizations where millisecond lags really actually change how something would perform in the marketplace, such as when you work with clients in things like manufacturing environments, computer vision environments, where you need to process things very fast,’ says Jim Rowan, Deloitte principal strategy and analytics practice leader.

IT consultant Deloitte has acquired an Israel-based solution provider that specializes in helping clients take advantage of GenAI opportunities on Nvidia-based infrastructure.

With the acquisition, Deloitte gains access to a team of developers with strong experience in the Nvidia CUDA parallel computing platform and programming model, said Jim Rowan, a Deloitte principal who leads the New York-based company’s strategy and analytics practice.

“It's one of those things where we you see an amazing company and get connected with them through the right channels,” Rowan (pictured) told CRN. “And then I think what they were offering in the high-performance compute space is really what we're going to be seeing a lot of going forward. So our team was able to get in front of them pretty early and excite them about joining a bigger firm and getting access to a lot of clients.”

Deloitte learned about OpTeamizer, which was founded in 2015, via its contacts in Israel who work with startup ecosystems, Rowan said.

OpTeamizer brings to Deloitte a very unique skillset, one that Deloitte is anxious to build on, he said.

“They work really closely with Nvidia’s CUDA stack,” he said. “They can help optimize AI models for running on Nvidia GPUs, where, in the market today, a lot of AI is running. Really, what they're doing is these high-performance optimizations where millisecond lags really actually change how something would perform in the marketplace, such as when you work with clients in things like manufacturing environments, computer vision environments, where you need to process things very fast.”

OpTeamizer also tackles efficient processing on the edge, Rowan said.

“So when you think about edge AI solutions, how do you optimize for that,” he said. “So we were intrigued by that because we really see the direction of travel in the AI space is terms of, yes, we have large language models, and they're important for what we do. But where do we see the small language models go? The privatized models that individual clients have pulled together? And also we're seeing what things can happen on a device. So we think direction and travel in this space closely aligns with some of the work that they do.”

While Deloitte already has a strong AI practice, OpTeamizer brings a unique CUDA experience, Rowan said.

“We have a huge number of practitioners that are trained, enabled, and certified on AI technologies,” he said. “What's unique about OpTeamizer is their CUDA experience. And it's both the experience they're bringing, but also their methodology for training and developing others to scale the practice. OpTeamizer Founder Tomer Gal is also an educator. He's got a great background about how to develop and build AI talent which I think is super important. And if you look at some of the survey data that we've done on the GenAI market and what clients are looking for, the big question is, do they have the talent to succeed in this space? Tomer is a great educator for us as we continue to scale our practices.”

Gal has already done a presentation at Deloitte’s AI forum in India, and is slated to do a presentation in mid-March at the Nvidia GTC in San Jose, Calif., Rowan said.

Deloitte’s acquisition of OpTeamizer closed in early January and was announced last week. Deloitte declined to discuss the financial terms of the deal. Rowan also declined to discuss OpTeamizer’s finances.