One phone screen then four back to back 1:1 behavioural/coding interviews. Each interview (including the phone screen) basically follwed the format of behavioural question (or two) then one or two coding questions. The coding questions usually weren't inherently difficult but the interviewer would ask how it scaled, how you might test it and generally get you to talk about your solution. It seemed like implementing a non-optimal solution was OK as long as you were doing it for simplicity, you understood why it was non-optimal and could explain how you might improve your solution. One note is that they called me back with decision the next morning. I was impressed with the speed of this part of the process.