Google Data Scientist Interview Experience — Two Rounds, Rejected Three Weeks Later

I got pulled back in and then had two rounds. Interviewed in early January.

The first round was coding. Three questions — I don't fully remember all the details.

One was something like: you're given a random generator that can produce a uniform sample in (0,1), and you need to write a function that returns a 2-D uniform sample in the square (-1,1) x (-1,1). Treat the random generator as a function you can just call.

Another was the longest increasing continuous subarray — I don't get why they're still testing this kind of LeetCode question.

The third one was something like this: you have a bunch of search queries split into a bunch of buckets (a1,b1,c1), (a2,b2,c2), ... where each bucket stores left_bd, right_bd, and search_count. There are K such buckets. You don't know the individual search counts, only the bucket boundaries, and you're asked to give an approximate value for the nth percentile.

At first I found the bucket the percentile falls into and said just return mean(left_bd, right_bd). He didn't seem too happy with that, and asked if I had a better method. I said linear interpolation within the bucket.

The second round was with the team lead. They just asked about the project on my resume — no coding. We talked for maybe half an hour, less even, and they said it was fine and left time for questions.

I got rejected about three weeks later.