Started with a phone interview with the current data scientist working there. The next day I was sent a coding challenge which I'd been told some of the details of during the phone conversation - two problems, choose one, no strict time limits. Emphasis was to be put on reusability and generality of code moreso than results, since I have a strong numerical background but less of a software background. I made an error in the calculation but it seemed to not be such a big deal. This would have been followed by an in person interview but we decided not to go ahead since I was not 100% on quitting my PhD.