What is the most important/painful step in NLP Data Management?

Web: https://www.reddit.com/r/LanguageTechnology/comments/vhj2g3/what_is_the_most_importantpainful_step_in_nlp/

Hi everyone!

I am doing research for a project regarding NLP Data Management.

My team and me identified the following five overarching building blocks in machine learning data management.

[https://imgur.com/a/JtMTzYS](https://imgur.com/a/JtMTzYS)

Now specifically in regard to NLP. Which one of these steps do you regard as most important / most painful?

I’d be really happy for any (gladly very specific) examples you encounter in your work or research.

Thanks in advance!

data data management languagetechnology management nlp