Skip to main content

Research at Unbabel

I am thrilled to initiate a series of posts describing some recent exciting news. At Unbabel, we have been pioneering advancements in machine learning since 2015 when we initiated Unbabel AI Research. Throughout the years, we have been pushing forward the state of the art in translation quality estimation, evaluation metrics, and automatic post-editing towards building the World’s Translation Layer

Our work has been presented at top conferences in the field (ACL, NAACL, EMNLP, ICML, NeurIPS, WMT, AMTA, EAMT). We have fuelled open-source projects such as OpenKiwi and COMET that are now relied upon by many researchers around the world in industry and academia for translation quality estimation and evaluation. You can find some of our work at 

With that, our team is putting pen to paper and providing a platform to give other researchers and engineers a glimpse into our latest advances in research at Unbabel, with the goal to inspire others to use and build their research on top of our work and to contribute to our open-source tools.

We will kickstart this new R&D blog with a sequence of posts showcasing several of our recent achievements in research with topics that include:

Stay tuned!

About the Author

André Martins is the Vice President of AI Research at Unbabel and an Associate Professor at Instituto Superior Técnico, University of Lisbon. His research, partially funded by an ERC Starting Grant, includes machine translation, quality estimation, structure, and interpretability in deep learning systems for NLP. André has published 100+ papers with over 6,000 citations and an h-index of 37 in top-tier conferences and journals. He co-founded and co-organizes the Lisbon Machine Learning School (LxMLS), and he is a Fellow of the ELLIS society.

Profile Photo of André Martins