Peter Latham was a physicist for a large portion of his career (undergrad, grad and postdoc), but in 1996, he switched to theoretical neuroscience. After one-and-a-half years at NIH and six at UCLA, in 2004 Latham landed a faculty job at the Gatsby Computational Neuroscience Unit at University College London, where he has been ever since. Latham is interested in understanding how biologically realistic networks carry out computations, with a focus on probabilistic inference. This work borrows heavily from fields such as neural coding and the dynamics of networks of spiking neurons.