Tech

Twitter A.I. guru who worked for Magic Pony is set to join Cambridge University

Key Points
  • Artificial intelligence expert Ferenc Huszár is set to leave his full-time role at Twitter and join the University of Cambridge.
  • Huszár was one of the first employees at London-based machine learning start-up Magic Pony, which was acquired by Twitter for a reported $150 million in 2016. 
Social media apps displayed on a smartphone.
Fabian Sommer | picture alliance | Getty Images

Artificial intelligence expert Ferenc Huszár is set to leave his full-time role at Twitter and join the University of Cambridge.

His departure comes four years after the micro-blogging platform acquired the London start-up he worked at, Magic Pony, for a reported $150 million.

Huszár, who is a senior machine learning engineer at Twitter's London office, is joining Cambridge's Department of Computer Science and Technology, where he completed a PhD in 2012. 

"I can confirm that I'm taking up a role at Cambridge," Huszár told CNBC via email. "This also means that I will not indeed be full-time employed by Twitter," he added. "I will miss the team and company very much."

Huszár, who was one of the first employees at Magic Pony, is expected to start at Cambridge on July 7. 

The university said Huszár has an "outstanding academic record and is a highly creative researcher with broad interests in many areas of machine learning." 

The move is a relatively rare example of an AI specialist moving from a big tech company to academia. Indeed, it normally happens the other way round. 

In recent years, U.S. tech giants have used large salaries to lure top professors away from universities around the world, leading to concerns of an academic brain drain. 

Some AI researchers at Elon Musk's OpenAI are on salaries in excess of $1 million, according to The New York Times

Video compression

Magic Pony developed technology to improve the quality of compressed videos so they may be enjoyed at higher quality when limited bandwidth is available. The software complemented existing compression technology but it didn't substitute it. 

At the time of the acquisition, Twitter's chief executive, Jack Dorsey, said Magic Pony's technology would be used to boost live and video offerings and that the company's technology "opens up a whole lot of exciting creative possibilities for Twitter."

At Cambridge, Huszár will work alongside Professor Neil Lawrence, who left his job at director of machine learning role at Amazon last year to take up a role at the centuries-old university that is funded by Alphabet-owned AI lab DeepMind.

"We're very excited to welcome Ferenc to Cambridge," Lawrence told CNBC. "He is a rare example of an individual who has deep insight into the theory of machine learning algorithms as well as practical experience of deploying these algorithms at scale."

"This mix of understanding is vital to ensure academic machine learning continues to be driven by the real challenges industry face in deploying safe, reliable, fair and interpretable machine learning."

Other new machine learning hires at Cambridge include Carl Henrik Ek, who is joining from the University of Bristol, and Nicholas Lane who is joining from the University of Oxford.

Twitter declined to comment when contacted by CNBC.