Data scientist interview @ Uber

Any tips on how to prepare for the on-site? Sample interview questions? Should I focus more on practicing coding problems (data structures etc.) or on ML and stats? For the latter, what is the level of details of the questions? Heavy on the theory/concepts, math/equations, or application to business problems? A few examples or sample questions would be very helpful. I come from an applied background and I don't have formal education in CS or ML. I do, however, have a good understanding of ML concepts and algorithms from self study and side projects. And I write code on daily basis at my current job (data manipulation) in SQL R, and Python.