Making credentialling like a sport

Some of you may know that Kaggle’s motto is “We’re making data science a sport.™”. Now we’re publishing a leaderboard of our top ten performers. And it’s quite an eye opener.  There’s not a professor there. Indeed there’s not a  person from a top university there. Just ten of the best data scientists in the world. Their names and bios below the fold. Of course there are only a few disciplines that can be reduced to the sporting formula to determine the best, but if there were more, then we could be on the cusp of a revolution. 

1. Alexander D’yakonov

An academic in the Faculty of Computational Mathematics and Cybernetics department at Moscow State University, Alexander modestly describes his favorite problem-solving technique as “luck.” Despite this, the 33-year-old Russian has earned a reputation for using methods known for their theoretical rigor and elegant simplicity. This helped him to win the dunnhumby Shopper Challenge, which asked competitors to predict the amount and timing of supermarket shoppers’ next spends.

2. Sergey Yurgenson

With a Ph.D. in physics from Leningrad (now St. Petersburg) State University and his current position as a research associate in neurobiology at Harvard Medical School, Sergey Yurgenson has combined both of his areas of expertise to develop computational algorithms inspired by biology. Favoring neural networks, a type of learning algorithm modeled on how brain cells work, the 50-year-old’s best finish came in NASA’s Mapping Dark Matter competition. To tackle the task of analyzing images of galaxies, Yurgenson combined several different kinds of neural networks.

3. Vivek Sharma

With a master’s degree in computer science from the University of Virginia, 32-year-old Vivek Sharma is now based in Delhi, where he has become one of Kaggle’s most consistent performers. His recent best results were in the credit scoring competition Give Me Some Credit and in the Algorithmic Trading Challenge.

4. Jose H. Solorzano

Jose H. Solorzano is a 42-year-old software engineer based in Quito, Ecuador. He has previously worked on open source projects, including the robotics software behind Lego Mindstorms. Solorzano’s software has been used to design a space junk collection system constructed entirely out of Lego pieces and put to test in orbit. His most notable Kaggle success was when he won Don’t Overfit!, a competition aimed at improving predictive algorithm strategies that attracted the attention of many of Kaggle’s most enthusiastic users.

5. Xavier Conort

Xavier Conort is a French actuary whose adventures have taken him to Brazil, China, and Singapore, where he is currently based. He is the founder of Gear Analytics, a consulting firm that helps insurance firms and other companies use predictive modeling. The 39-year-old’s affinity for American car culture helped him win the hotly contested Don’t Get Kicked! competition, which asked Kaggle contestants to develop a mathematical model that can work out which used cars are most likely to be bad buys, or “kicks.”

6. Tim Salimans

Tim Salimans is a 26-year-old Ph.D. candidate in econometrics at Erasmus University Rotterdam in the Netherlands. He triumphed in the World Chess Federation (FIDE)-sponsored Deloitte/FIDE Chess Rating Challenge on Kaggle. The task was to predict the outcomes of chess games, thus developing a better ratings system. Saliman’s trick was to create a modified version of the method used by Microsoft to rate Xbox players, a variant that ended up being better than standard ranking algorithm.

7. Vladimir Nikulin

Vladimir Nikulin has a Ph.D. in mathematical statistics from Moscow State University and has worked as an academic in Russia and Australia. A veteran of data mining competitions, the 52-year-old sees them as an essential part of the research process since they help researchers identify real from illusory progress in their methods. His best Kaggle finish was third place in the used car defect prediction challenge ‘Don’t Get Kicked!’.

8. David J. Slate

David J. Slate holds more than half a century’s experience in programming. He was building operating systems and cracking chess programs decades ago, winning the World Computer Chess Championship in 1977. A former academic at Northwestern University in Evanston, Ill., today he spends his retirement entering Kaggle competitions under the moniker “Old Dogs with New Tricks.” His best performance on Kaggle came when he won the R Package Recommendation Engine competition for recommending software packages to users in the programming language R.

9. Yannis Sismanis

Yannis Sismanis is a researcher at IBM’s Almaden Research Center in San Jose, Calif., where he works on data intensive analytics. Originally from Greece, his background is in electrical engineering and computer science. Sismanis won the first of two Kaggle competitions aimed at improving chess ratings systems, called Chess ratings – Elo versus the Rest of the World.

10. Jason Tigg

Jason Tigg grew up on the Isle of Thanet in southeast England and goes by the alias “PlanetThanet” on Kaggle. At age 14, he used assembly language to build a program that could play Othello. The 43-year-old now holds a Ph.D. in elementary particle physics from Oxford and is based in London, where he works in the finance sector. Tigg’s performances on Kaggle have been outstanding, notably in the Photo Quality Prediction competition and the Claim Prediction Challenge.

2 thoughts on “Making credentialling like a sport

  1. Most of them have PhD’s from good Universities like Moscow, Leningrad and Erasmus. Jason Tigg has a PhD from Oxford c.f. your comment “Indeed there’s not a person from a top university there.” I guess you meant “there’s not a professor from a top university there” but then what about #1 and #2? They are academics.

    Why would you expect academics on the list anyway? Academics think about new ways of framing and analysing problems. I would not be able to justify spending time on Kaggle problems. It brings no benefit to my workplace so I would have to do it in my own time. Having a family makes this difficult. I wonder how many of these top 10 are well-balanced individuals! They are sure gender skewed.

    Kaggle is a great way of harnessing nerd-power into something more constructive than trojans and spam. The list is pretty close to what I would have expected.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Notify me of followup comments via e-mail. You can also subscribe without commenting.