At Aggregate Knowledge, Matt is in charge of "making the science work." This entails algorithm development for the massive data volumes at AK, wrangling the science team, and thought leadership for attribution modeling. Matt has been a startup scientist for 10 years, working in diverse fields such as machine learning, recommender systems, and simulation. His current interests are in very large data set manipulation and probabilistic algorithms. Matt received his M.S. from Brandeis University in computational neuroscience and a degree in Physics from Reed College.