Overview

NLG has been designing and implementing in-house hybrid machine translation (MT) systems, which synergize the benefits of both statistical machine translation and rule-based machine translation, since 2014.

Our ensemble architecture integrates the latest state-of-the-art technologies (Moses, KenLM, MGIZA++) and linguistic engineering techniques in conjunction with high quality client-specific bilingual corpora and term bases to generate the highest quality output.

Our expert team of engineers and linguists combine extensive knowledge with a scientific approach to MT and post-editing of MT output. This allows NLG to identify robust methodologies & workflows and provide translators with best practices, guidelines, and continuing training.