English is by and large considered as the lingua franca of scholarly communication. Such a generalised use has certainly the advantage of facilitating international exchanges, but it also generates inequalities among researchers, and limits the dissemination of scientific knowledge. Although translation could be promptly identified as the solution, scholarly communication has historically been marked by a shortage of human and financial resources to support traditional translation processes. The goal of this paper is to present a multi-user approach to machine translation evaluation for a use in scholarly communication. In particular, the paper introduces the fine-tuning and evaluation methodology set up to comply with the needs of different target user personas (translators, researchers, readers). Given the focus of the conference, the paper will describe in more detail the evaluation methodology related to the “Translator” persona. The paper will also include general preliminary conclusions, and information about the on-going evaluation work.