PhD student reaches Google Code Jam championship finals

By Jean Joyce

Computer Science graduate student Anders Jonsson competed against more than 7,500 participants from more than 100 countries to reach the championship finals in Google''s Code Jam 2004 computer programming competition. As one of the 50 finalists, Jonsson was flown to Google headquarters in Mountain View, Calif. to compete in the championship round on Oct. 15.

"I was very happy just to qualify for the final round since I didn''t have a lot of experience with these competitions," said Jonsson. "More than the competition itself, I enjoyed meeting the other competitors from all over the world and visiting Google as it is one of the quickest growing companies at the moment."

The competition began Sept. 1 with an initial qualification round, and 500 of those went on to a two-round competition field. The top 50 scorers from round two of this phase went to Google for the finals. The 50 finalists who are working or studying in the United States and in 16 other countries, from Scandinavia to central Europe to Hong Kong, Korea, Australia and New Zealand, competed for $50,000 in prize money. All of the programming for any round could be done in Java, C++, C# or VB.NET.

"The Google Code Jam is one way Google encourages and supports the engineering, programming and computer science communities around the world," said Alan Eustace, vice president, Engineering, Research and Systems Lab, Google Inc. "We’re continually exploring new opportunities to reach out to smart, talented people who enjoy solving problems. This is a fun way of finding, rewarding and potentially recruiting some of those people to Google."

This is the second year of the Google Code Jam, which is produced in conjunction with TopCoder, the leader in online programming competition.

Jonsson is a graduate student in the Autonomous Learning Laboratory. He is expected to graduate with a Ph.D. in September 2005. Jonsson is advised by professor Andrew Barto, and his dissertation is titled "A Novel Approach to Abstraction Discovery in Reinforcement Learning."