Vancouver, BCCanada V6B 0J2
April 3, 2014
Yes and previously yes.
Whatever is new and exciting. Always Convex-Op, Boyd and Vandenberg's books are my bible. B-matching is a particular area I am exploring based from Jebara's interesting work with loopy belief propagation. I enjoy hiring and learning from other ML folks. computational complexity, How to make small data sets (<1m labeled items) work with machine learning via SSL. Efficient feature selection. Measuring the predictive value of a feature given a non linear correlation with other features. Live and breathe this art form.
CEO of Steamsharp, my company applies ML algos amongst other numerical methods to large sports data. We create market making technology for sports betting conglomerates worldwide.