My research, under the supervision of Dr Ivan Titov and Dr Alexandra Birch, is focused on low-resource machine translation, i.e. on creating models of (neural) machine translation that can thrive under data sparsity. This research is part of the GoURMET project, a European collaboration between academia and media to advance journalism to and from low-resource languages. In a broader sense, I am interested in deep generative models and variational inference, especially when applied to natural language processing. One day, I hope we can approach models that can generate convincing natural language from true language understanding.