I did my PhD in Informatics at Edinburgh University (UK) with Chris Williams. I am a fan of guess-what comix. I moved on to sunny California to do a postdoc at UC Berkeley with Michael Jordan and Peter Bartlett. Currently, I work in Bernhard Schoelkopf's group at the Max Planck Institute for Biological Cybernetics in Tuebingen, Germany.

I am interested in Gaussian process models and other non-parametric methods, PAC-Bayesian bounding techniques, support vector machines, Bayesian learning, adaptive metrics, variational approximations to Bayesian inference, supervised learning with partially unlabeled data, and convex optimization. Recently I got interested in temporal and spatiotemporal models for neuronal data, in and Bayesian experimental design (with applications to Bayesian gene regulatory network identification, undersampled fMRI reconstruction, and compressed sensing).