I think it's too early to say, because the impact of work is seen over time.

But candidates probably include Breiman/Cutler for Random Forests, Dean/Ghemawat for MapReduce, Peter Shor for Shor's algorithm, Daphne Koller for machine learning, Jon Kleinberg for network algorithms, Luis von Ahn for human computation.