I had the on-site interview in Zurich, Switzerland. It consisted of 3+2 rounds of 45 minutes each, with a short lunch break in between.

Four interviews focused mostly on knowledge of data structures and algorithmic thinking, whereas the fifth revolved around system design. I found only 2 of the algorithmic problems particularly difficult; the other 2 were fairly straightforward, but I guess this pretty much depends on the type of problems one has previously been exposed to.

The system design interview required prior understanding of cloud-based systems, which I, as a software engineer who mostly worked for automotive and semiconductor companies, did not really have. However, this did not prove to be a dealbreaker in the end. I ended up below the bar mostly due to not being fast enough in solving the algorithmic problems and not being knowledgeable enough about some of the more advanced data structures.

My experience was fairly positive overall. I especially appreciated the fact that Google provides interviewees with detailed feedback on their performance and advises them on what areas to focus for a possible re-run in 12 months' time. I don't remember any other company I've ever had contact with doing this, so a big thumbs up for this. They also communicated very well and professionally throughout the whole process; the requirements and timescale for every step were always made crystal clear, and they offered time for preparation.

One aspect I found problematic, though, was the lack of breaks between the interview rounds. I would have really appreciated a 5-10 minute break before delving into a new problem, just to have a chance to clear my mind. Going for 6 hours without even a 5-minute break can be painfully exhausting and it makes an already stressful situation more demanding than it should reasonably be.

Another issue I have with Google's interview process is that it is somewhat skewed toward fresh graduates and places experienced developers at a disadvantage. Ten years have passed since I last needed to implement binary search or Dijkstra's algorithm, whereas a fresh grad probably had an Algorithms & Data Structures course sometime last year. Sure, you can always refresh your memory and relearn all the algorithms you have forgotten in years of industry work, but it makes the required effort much greater for experienced developers, especially if you have to prepare for the interview while having a full-time job and a family.

How would you design a cloud-based photo storage and sharing application with a smartphone frontend?