If you're a Python coder who wants to work for a hedge fund and you don't know Wes McKinney, you might want to familiarize yourself with the name. McKinney started working for Two Sigma, the huge quant hedge fund with around $40bn under management, in August 2016. He's best known, however, in the world of open source programming.
An MIT math graduate, McKinney began his finance career at hedge fund AQR Capital Management in 2007. Three years later, he quit to do a PhD at Duke. Since then, he's been steeped in the world of Python data analytics tools, including Datapad, a visual analytics system he sold in 2014, and the open source Pandas Python data analysis library, which is nine years old this April.
"Open source is the future of how a lot of our software will get built," said McKinney, speaking at a Newsweek conference on artificial intelligence and data science in London today. However, because open source software is typically developed by engineers working on unpaid side projects, he said many open source initiatives are unsustainable and can absorb all engineers' free time. Open source programmers, "have lives and families," said McKinney, adding that he personally has started limiting his involvement for this reason. "I personally feel that if my open source work isn't being sponsored, I ought to focus on things that can pay my rent."
McKinney didn't say so, but this might be why he joined Two Sigma, which pays an average of $417k to each of its staff in London. Two Sigma, along with rival quant fund Man AHL, is one of the pioneers in the use of open source code in trading applications. "The best way for finance to contribute to open source is to dedicate engineering time at the organization to allow open source projects to survive," McKinney said. This way, open source coders will get paid for their passion.
For this to happen, however, finance needs to open up. Ever-secretive, financial services firms have traditionally been mistrustful of the open-source communities where everything is shared by definition. This is changing, but only on the periphery. Gary Collier, co-chief technology officer at hedge fund Man AHL, said AHL doesn't use open source code for the proprietary algorithms in its trading systems, but for the "building blocks". Airflow, a tool developed by Airbnb, is being trialled as a scheduling device by the fund, for example.
As hedge funds wake up to the importance of open source, involvement in communities like Github is a valuable source of CV points for potential recruits. "If you want to recruit top talent into an organization, the organization needs to get into open source," says Collier. Equally, candidates need to show evidence of involvement in open source projects. "If you're a techie, you need to be on Github," says another senior hedge fund technologist. "It's like being an artist or an architect - you need your side projects, you need to point people towards your portfolio."
This is all well and good, except developing your open source portfolio takes time. And as McKinney points out, time is short when you've got a day job and family to consider. It's even worse if your open source project takes off. In this case, McKinney said you can become "overburdened with success." For example, his Panda system - which has 9,000 stars on Github - reached a point where it required so much maintenance and stabilization time that he had little time for innovation or anything else. "It becomes a full time job," said McKinney. "Which is difficult when people are working on these things on a voluntary basis."