The development of technology, in particular, innovations in natural language processing and means to explore big data, has influenced different aspects in the training of translators and interpreters. This paper investigates how learner corpora and their research contribute to the teaching and learning of translation and interpreting. It starts with a review of the evolvement of learner corpora in translator and interpreter training. Drawing on data from the Chinese/English Translation and Interpreting Learner Corpus (CETILC), a learner corpus developed for the study of lexical cohesion, the paper introduces three case studies to illustrate the possibilities of exploring learner data through human annotation, machine-facilitated human annotation, and finally human-supervised/edited machine annotation. The findings of the case studies suggest the complexity of learner language and its intricate relationships with various factors concerning the learner, text, and task. The paper ends with a discussion of the great potentials of purposely made learner corpora such as the CETILC in translator and interpreter training, as well as the application of learner corpora in (semi-) automatic processing of learner texts.