Venture capital funding is rushing into machine learning startups

At just 30 years old, Shivon Zilis is making important deals at Bloomberg Beta that are shaping the future of the machine learning business

Shivon Zilis

Shivon Zilis. (Portrait by Roberto Caruso)

Several years ago, Big Data was the new thing turning every venture capitalist on his (or, only very occasionally, her) ear. Today’s go-to trend is using machine intelligence (MI) to help businesses exploit their data to work smarter. “In the next five to 10 years, it’s going to be very difficult to see any data-oriented piece of the world not fundamentally transformed by this technology,” predicts Shivon Zilis, an investor at Bloomberg Beta, based in San Francisco and New York.

Machine intelligence startups have amassed almost US$1 billion in investments since the technology began to look viable in 2010, according to CB Insights. The rapid development in this nascent field has spawned venture capital funds, like Data Collective and Bloomberg Beta, that invest with a focus on MI startups. Zilis, who just turned 30 this year— she appeared in Forbes magazine’s “30 Under 30: The Top Young Investors of Venture Capital” in 2015—has co-led between 15 and 20 of the 68 deals Bloomberg Beta has closed since its founding in 2013. A number of those investments have gone on to raise capital at higher valuations, according to Forbes. One of them was Newsle, a real-time social media notification system, which was acquired by LinkedIn Corp. (NYSE: LNKD) in 2014 for an undisclosed amount.

A proud Canadian, Zilis left her Toronto home after winning both a hockey and an academic scholarship to study in the U.S. She picked the latter and completed her bachelor’s degree in economics and philosophy at Yale. “I was obsessed with technology back in school but wanted to be on the business side of technology,” says the self-described “data geek.” She first went to work for IBM Corp.’s (NYSE: IBM) global microfinance initiative, where she helped create solutions to enable loans to be made to small businesses in developing countries. Her next job had her advising startups in the education and data services spaces for Bloomberg’s internal incubator. It was there that Zilis and her colleagues saw an opportunity for a traditional venture capital fund with business information giant Bloomberg LP as the limited partner. The $75-million Bloomberg Beta fund was created to back early-stage technology companies with the potential to transform the future of work.

Currently, the majority of Bloomberg Beta’s investments are in MI companies based in the U.S., though three are headquartered in Canada. The six-member investment team has examined nearly every type of MI startup—more than 2,000 companies in the enterprise, agriculture, healthcare, transportation and logistics sectors—but Zilis is picky about where she puts her money. For her, investing in startups is about investing in people who have the whole package. “In the MI space, it’s really hard to find a mix between extreme technical competence and a deep personal desire to solve a problem,” she explains.

One company that met her criteria is Toronto-based Deep Genomics, which applies machine learning to early disease detection to make possible more targeted therapies. Zilis was impressed by founder Brendan Frey’s technical ability and commitment to the problem. Bloomberg Beta and True Ventures together put US$3.7 million into the company last November.

Bloomberg Beta works with founders to build their companies in a sustainable way. “We encourage them to be very thoughtful in how they spend their money and the rate [at which] they make hires,” says Zilis. The venture fund is also disciplined with its own investments. The team focuses on areas where its members believe real problems are being solved; the investors try to avoid getting pulled into the hype cycle. “A lot of our companies are selling specific enterprise contracts,” Zilis points out. “As a result, they’ll be better protected from fluctuations in the startup ecosystem.”