Using a large training dataset of highly relevant, quality-controlled translations of news content, we were able to successfully build English-vernacular machine translation models for many low-resource languages.
We continuously build machine translation models and improve our translation capabilities, including audio and video to text.