How would you build a model to predict someone's Twitter username from their name, age, geography, and other publicly available data on their Twitter profile?

Got asked this question during a data science internship interview. Besides the obvious fact that someone's Twitter username is public, how would you answer this?

I initially said I would use a supervised learning model, the interviewer then said what if the project was resource constrained so we wouldn't be able to label the data. I then said I would probably use some kind of sub-string selection using a concatenated string of their profile features.

data datascience geography profile twitter