Google Senior Data Scientist interview questions

This was a conversion to FTE interview process as I was previously working as a contractor. After talking with a recruiter in HR to get some basic info, I interviewed with several within leadership of the group I was already working with. I was told shortly after this interview that the level of the role may be a mismatch for my experience (the number of years worked). I declined to proceed with the process, because a reduction in level would also result in a decrease from my current salary.

I was contacted by a recruiter who got my email from my PhD department's website. I had a phone screen with the recruiter who then set up a technical call for me. They said that there is either one or two technical calls before onsite depending on how well the first one goes. The technical call was not too tough - we talked through a relatively simple hypothesis testing problem for proportions. There was also a pretty basic SQL question. They invited me onsite to Mountain View for the final round. I was applying for a job in NYC, and lived in NYC, so I found this kind of surprising but they said they do all their interviewing in CA. When I arrived on the day of the interview, I was REALLY surprised to find that three out of five of my interviews were over videochat. The interviewers were on different parts of the Google campus, and they didn't want the inconvenience of walking to meet me in person (in their defense the campus is pretty big). But this seemed kind of crazy in light of me taking an almost two day trip in order to visit. This definitely colored my overall take of the company and how they treat their candidates. Nobody apologized to me for this or noted that it was strange at all. On top of that, the tech was not 100% functional. Some of the cameras were old and/or not working, leading to unnecessarily long prep before interview actually began The actual questions were pretty tough. Most were more statistical than programming-based. They were interested in my understanding of parameter estimation for click through rate problems and also my knowledge of linear models with and without regularization (ie lasso and ridge regression). The questions seemed tough but fair. People were generally pretty nice. The last interview of the day was behavioral. It was hauntingly corporate-feeling and not what I expected from Google: "Describe a time when ..." "What is your greatest weakness" etc. The campus in Mountain View pretty cool. It feels like you're on a college campus. It took about 3 weeks post-onsite to get their decision.