In a previous article, I’ve told you the story of how we fixed the problem of sentence tokenization in RTL languages. If it’s only this specific technical problem you were interested in, read no further! I’m not going into more detail about the technical solution.