U.S. Markets open in 4 hrs 20 mins

Little-Known Data Wizards Take On the Pros at New AI Hedge Fund

Justina Lee

(Bloomberg) -- The first question Mike Merritt-Holmes gets asked when pitching his AI strategy is about his resume: which hedge fund did he work for? But Kvasir Technologies Ltd. doesn’t come from an established corner of finance. And neither does the man touting it.

Merritt-Holmes founded the machine-learning fund with two fellow computer scientists who made their mark in the world of big data at analytics firm Teradata Corp. The trio subsequently harnessed their engineering acumen to power machine-learning stock investing in developed markets, financed by their personal savings.

Before they crossed paths, only one had professional market experience. Emboldened by their trading successes, the team now expects to oversee around $135 million by December.

Kvasir is among a cohort riding a wave of interest in artificial intelligence to uncover new investing riches -- empowering computer nerds and data wizards to take on seasoned money managers on Wall Street.

“People are really keen to see fresh new thinking and fresh new teams that aren’t coming from a regurgitated reinvention of previous funds,” said chief operating officer Merritt-Holmes from his WeWork offices in London.

Machine learning in finance is a broad term for techniques ranging from complex statistical analysis to parsing reams of text. It’s increasingly where the action is. Data provider Preqin Ltd. tracks 305 AI hedge funds overseeing a total of $17 billion, according to an Aug. 30 note.

The technology has plenty of skeptics citing noisy security prices, with the underperformance of a Eurekahedge AI index in the last three years underscoring the challenge of using machines in practice.

That’s not stopping the tech crowd. U.S.-based AI hedge fund FORA Capital, for example, is led by entrepreneur Stanislav Shalunov, who’s worked at BitTorrent and the firm behind the FireChat app. Software engineers have a prominent presence at Walnut Investments, an AI fund in Paris, which counts trading-technology expert Manoj Narang among its advisers, according to its website.

Before Kvasir, Merritt-Holmes established a data startup that was eventually sold to Teradata Corp. Co-founder Mads Ingwar has a Ph.D. in machine learning, and CIO Martin Oberhuber is a former quant at high-frequency trading firm RGM Advisors.

Kvasir started trading its main fund in August. Its clout has been enhanced after advising industry giant Brevan Howard Asset Management this year on an AI allocation. The hedge fund declined to comment.

Billionaire Alan Howard has long harbored an interest in the technology and has previously invested in AI firms.

No Autopilot

Kvasir’s founders call theirs an end-to-end approach, using automated programs to process data, construct portfolios and control risk. But they emphasize it’s not a robot spitting out trading signals on autopilot. “We usually always start out with an economic intuition and we design the framework of the signal,” said Oberhuber. “Then we let the machine-learning algo do its own thing.”

The firm, currently a team of just five, is authorized by regulators through G10 Capital Ltd. which provides services to fledgling funds. Kvasir owes its moniker to an omniscient creature in Norse mythology who travels to spread knowledge -- but ends up killed.

“Let’s hope that doesn’t happen,” said Merritt-Holmes.

(Adds detail on performance for AI index in seventh paragraph)

--With assistance from Nishant Kumar.

To contact the reporter on this story: Justina Lee in London at jlee1489@bloomberg.net

To contact the editors responsible for this story: Samuel Potter at spotter33@bloomberg.net, Sid Verma

For more articles like this, please visit us at bloomberg.com

©2019 Bloomberg L.P.