Researchers working in machine translation have benefited from the availability of large-scale corpora, and as a result in recent years an increasing number of empirical methods have been proposed. This chapter presents a brief overview of EXPERT (EXPloiting Empirical appRoaches to Translation), an FP7 EC-funded project whose main aim was to promote the research, development and use of datadriven hybrid language translation technology. Given the importance of translation memories in the everyday activities of professional translators, the chapter presents three research directions pursued in EXPERT which aimed to develop data-driven tools that are directly useful for translators.