Ekaterina (Katya) Garmash

I'm a Research Engineer at Unbabel working on machine translation and related tasks. Before joining Unbabel I worked as a data scientist at KLM Royal Dutch Airlines. 

Before that I did a PhD at the University of Amsterdam (ILPS). My thesis was about statistical and neural machine translation and I was supervised by Christof Monz and Maarten de Rijke.  During my PhD studies, I interned at the NLP group at Amazon, Berlin and  at the machine translation team at Microsoft Research, Redmond. 

The topics I  have worked on include: syntactic language modeling for machine translation, multi-source neural machine translation, machine translation for morphologically rich languages, conversational modeling, text classification.


Contact: katya.garmash@gmail.com, Linkedin


Publications

  • Garmash, Ekaterina. Exploring the correspondence between languages for machine translation. PhD thesis. University of Amsterdam, 2017.  PDF 
  • Garmash, Ekaterina and Monz, Christof. Ensemble learning for Multi-Source Neural Machine Translation. Coling, Osaka, Japan, 2016. PDF
  • Garmash, Ekaterina and Monz, Christof. Bilingual structured language models for statistical machine translation. EMNLP, Lisbon, Portugal, 2015. PDF
  • Garmash, Ekaterina and Monz, Christof. Dependency-based bilingual language models for reordering in statistical machine translation. EMNLP, Doha, Qatar, 2014. PDF
For more information, see my Google Scholar profile.