Global hedge fund manager Citadel and market maker Citadel Securities have a mainline into the hottest data science talent through their iconic Data Open series, which will see 20 teams of students meet for the final championship at the New York Stock Exchange in November.
The competitions have been taking place in major centres across the UK, Ireland, Canada and US including Oxford, MIT, U.C. Berkeley, Harvard, University College Dublin and Columbia University, and have attracted thousands of PhDs, Masters students and undergraduates to compete for $25,000 worth of prizes and a chance to participate in the final championship.
The winning team at the "The Data Open Championship" at the NYSE will receive $100,000 prize, presented by Citadel CEO Ken Griffin.
Citadel's Head of Talent Strategy, Justin Pinchback explained that at each datathon, groups, usually made up of four students, spend a day working on a very large and very complex dataset. He said all the events have been oversubscribed and that typically there are about 400 applicants for 100 places.
The first datathon in the UK took place at Oxford University in October. "We saw some of the top minds from schools such as Oxford, Cambridge and Imperial go head to head and really practise the craft of data science," said Pinchback.
The students receive the dataset at the beginning of the day and then spend most of the day cleaning, analysing and gleaning insights from the data. At the end of the day, they submit their findings, usually in the form of a technical paper with some data visualisations, overviews and syntheses. A panel of judges then determine the winning submissions, as well as the runners up.
The datasets are not financial, but cover a wide range of topics. This levels the playing field by allowing students with non-financial or economic backgrounds to participate.
Pinchback said: "We have provided datasets about housing, income, transportation, Uber rider and taxi rider information. We think it is important to provide data that is relevant to everyone. We also find that data with high dimensionality best represents the type of data that employees can expect to work with at a firm like Citadel or Citadel Securities."
"It's very rare in the real world to see data that is nicely packaged and clean. A large portion of a data scientist's time is spent making sense of incomplete information that is unstructured and unorganised."
"We believe that the datathon program can help us obtain a good understanding of how people attack large, complex problems by offering them a very broad set of information to work through."
It's no secret on Wall Street that there's a shortage of data science experts, and getting in touch with the next wave of talent in this field is a tactical move.
"This has been a great way for Citadel and Citadel Securities to get in touch with this community," said Pinchback, "and while we have had very good recruiting outcomes, we are also trying to build relationships with many of these individuals that will last for years."
Newsweek's AI and Data Science in Capital Markets conference on December 6-7 in New York is the most important gathering of experts in Artificial Intelligence and Machine Learning in trading. Join us for two days of talks, workshops and networking sessions with key industry players.