Length 1 hour. 2 technical tasks in jupyter about pandas (efficiency will be scored more than other dev aspects) 2 questions about certain analytic metric and approaches to increase it. Interviewer was asking goals during technical tasks, but in feedback treated every your guess and his own reactions as a minus. Thus I recommend to focus on code itself and discuss only after this points are done. This is not a discussion format to fit code to interviewers tastes and expectations in live, but scored as solo performance, where you have to push most optimal solution and interviewer during first 2 tasks just have to control that you are not getting help of 3rd persons . At the end of interview jupyter notebook will be saved so I recommend to fix all your critical points and answers there, may be it will help further if interviewer misinterpretation will occur.