Rounds 1 and 2 were pretty smooth as they were Online Assessments. The nice thing about these 2 rounds was that candidates were given a one-week window by when they were supposed to complete the assessments, which meant the deadline for completion was more relaxed. Round 1 consisted of debugging 7 programs in 20 minutes and solving 24 Logic and Reasoning questions in 35 minutes. Round 2 consisted of solving 2 programming questions in 70 minutes. The final hurdle was the Virtual Interview round which is said to last for 45 minutes. I was asked a tree-based path sum question that although at first I suggested a naive approach which was quickly shot down, I came up with a valid workaround for it that even my interviewer seemed to like. I was asked about a few edges cases which I incorporated into my solution and once my interviewer was satisfied, I was asked to put the algorithm into code. I had the algorithm coded as a function within 5-10 minutes. My interviewer, despite conceding that my implementation was a correct interpretation of the suggested algorithm, went on to point out certain redundancies in my code with unnecessary if-blocks. On agreeing with him and getting rid of them, he was satisfied. However, by this time, I was aware that the interview had gone overtime and he just asked me if I had any questions for him. Despite getting the solution and 'up to the standards' of my interviewer, I received a swift rejection email within 24 hours. On asking a few of my peers ow their interviews went, I was shocked to know that some of them were asked ridiculously easy questions like printing out a multiplication table or printing out all odd numbers less than some x. The serious irregularity of the interview process and inconsistent difficulty level was a huge disappointment from a company of Amazon's stature.