BigDataFr recommends: What are 20 questions to detect fake data scientists?
Identifying the essential skills for data scientists.

I’ll stay away from code examples myself as they seem rather shop-specific and thus best designed locally, but if you want some questions, here you go. These questions are intentionally difficult and are more on the statistics/modeling side than the data processing side. That’s important, but someone else would be better poised to write those questions.

You might want « I don’t know, but what I would do is read the following sources…. » to be part of your accepted answers, as that’s partly testing honesty and forthrightness of the candidate. The last thing an organization needs is bullshit artists who overpromise what they can do or just make things up.

You really want to be wary of wanting the « unicorn hire. » […]

Read more
By Jay Verkuilen, PhD Psychometrics UIUC 2007
Source: quora.com