Making sense of large corpus of text via machine learning, say .. SVM which stands on top of linear algebra (computing inner products in an isomorphic space of bigger dimension via kernel functions) (Hillary Mason working at bit.ly gives talks about this stuff http://www.hilarymason.com/ )

Data mining, a lot of which boils down to multidimensional exploratory statistics needs linear algebra.