Internship consisted of a 3 months period sped with the Deep Learning group of Philips’ Data Science Department. The goal was to do low latency (6 seconds) sleep stage classification of raw EEG data. The approach we using Convolutional layers for recognizing features in the raw signal and then processing these features with Recurrent layers such that the network could grasp a sense of how the signal progresses in time. This approach yielded a state of the art performance on this type of data.