I have been following a lot of feedback on Google Interview experiences posted by other LC members and it seems that the pattern is well-known:

1. Phone Interview = 1 or 2 coding problem
2. Onsite = ~3+ coding + ~1 behavioral + ~1+ system design

I also noticed that for graduates most of the times there is no system design, but apart from that I don't see the difference on the structure.

So my first question is how does a Google Interview differs in the following scenarios:

A) from SE Level 4 to SE Level 5?
or
B) from applying to a SE Android position or a SE Machine Learning position or a SE Manager position?

My assumptions for A) are that the higher the level you apply the harder are the questions and also the higher is the bar to judge your performance relative to others? But since the problem-solving questions are usually abstract, how they access your background experience for instance, from Android to Machine Learning to Manager? I assume that probably would be on the behavioral interview where they fact check your CV with you?

Let me know what is your understanding on this...