The company takes a shotgun approach to recruiting, interviewing many more people than a typical company would for a role. They weed people out based on who can best spew code in 45 minutes for an inane problem that in no way represents anything anyone would ever code. The code is then evaluated by a machine learning system.

For the interview, they pair you with an actual person, but that person, in my case, didn't even know the language associated with the role I was interviewing for.

The company does not want unique, creative people. The company wants code monkeys.

Write code to represent a safe deposit box.

The problem with this question is that the U/X and the interactions with the "real world" are what matter, but the interview could not answer much about those aspects.

The code itself is nonsense absent a clear and complete analysis of U/X.