Veteran Durham entrepreneur Robbie Allen, left, is building a business based on machine-learning technology developed by Duke University Professor Lawrence Carin, right. Mike Salvino, center, a private-equity executive and a former unit leader at Accenture, is executive chairman of Infinia ML.
By David Ranii
Having hit it big with the reported $80 million sale of his Durham-based artificial-intelligence company Automated Insights Inc. in 2015, Robbie Allen recognizes that even more is expected of him in his new role as CEO of Infinia ML. The Bull City-based company, spun out of Duke University, deploys machine learning that enables its big corporate customers to reduce costs or boost revenue.
“The first time around, there were no expectations because I had never done it before,” says Allen, 41. “This time it’s different. I feel that I’ve been given raw materials that are very hard to come by … on a silver platter. If anything, I feel more pressure because I don’t want to screw this up.”
The tales of the two startups led by Allen are starkly different. Automated Insights, which provides automated news stories for Associated Press and Yahoo!, was sold to an Austin-based private-equity firm in 2015. It wasn’t a snap: Allen spent three years and $100,000 nurturing the technology from scratch before raising $1.3 million in seed capital. That enabled him to leave his job as a distinguished engineer at Cisco Systems Inc. and devote all of his time to the venture.
By contrast, when he joined Durham-based Infinia in November, the business had already raised $10 million from San Francisco private-equity firm Carrick Capital Partners, which co-founded the company just two months earlier. The company possessed a library of machine-learning technology amassed by a team led by co-founder Lawrence Carin, vice provost for research and professor of electrical and computer engineering at Duke University.
At Automated Insights, Allen was both the technology and the business guru. Now, when he and Carin meet with customers, he says, “it is maybe best I don’t talk. Larry is going to say it better than me. He knows way more than I do.” Carin has published more than 350 scientific papers and is, in Allen’s words, “a luminary” in the machine-learning field.
On the business side, Allen leans on Executive Chairman Mike Salvino, a Carrick Capital managing director based in Charlotte who spearheaded Infinia’s formation. Salvino previously was group chief executive at a unit of Ireland-based management consultant Accenture that has more than $7 billion in annual revenue and 100,000 employees. Salvino “has run larger organizations than 99% of the CEOs out there,” Allen says.
Allen’s own accomplishments include two master’s degrees from Massachusetts Institute of Technology, a half-dozen patents and authorship of eight technical books. “We couldn’t have hired anybody better,” Salvino says.
After stepping down as CEO of Automated Insights in April 2017, Allen’s idea of a break was to pursue a Ph.D. in computer science at UNC Chapel Hill. He was enjoying the program so much that he showed little interest when a recruiter approached him about the Infinia job. He changed his mind after meeting Carin and Salvino. “They blew me away, in a positive way,” he says.
Machine learning is a subset of artificial intelligence focused on creating algorithms that enable computers to extrapolate predictions from data — and continue to learn with new inputs. The technology is in such demand, and machine-learning experts are in such short supply, that corporate America has a 10-year backlog of projects, Allen says.
So far, Infinia has completed eight high-level machine-learning projects for Fortune 500 companies and is working on 10 others. The going rate per project has been “from the low six figures to the low seven figures.”
Projects have ranged from image analysis to upgrade the efficacy of X-ray scanners at security checkpoints to analyzing gene-expression data to determine the severity of cancer or to identify maladies.
Salvino, who had engaged Carin for a series of machine-learning projects at Accenture, got the Infinia ball rolling when he approached the scientist about forming a company. Carin had co-founded a Durham company, Signal Innovations Group, that was sold for an undisclosed price to BAE Systems in 2014. He wasn’t looking to start another business, but he recognized that machine-learning technology had matured. “The opportunity was profound. Carrick wanted to do it. It just seemed like the right time,” he says.
When Allen came on board in November, Infinia had seven or eight employees. By April 1, the staff had expanded to 15, seven of whom have Ph.D.s, and Allen anticipates a staff of more than 30 workers by the end of the year. For potential employees, the hiring process includes a 10-minute presentation about themselves; a 20-minute presentation on a project they’ve worked on; and a technical interview in which candidates identify algorithms and parameters needed to solve a problem.
Allen’s people skills sent him to the top when Infinia made a national CEO search. “In this business, if you lose the people, you’re going to lose the business,” Salvino says. “If you can’t build a good culture, this isn’t going to work.”