First, I had a phone screening with a recruiter. As usual, I was asked about my past experiences and a few questions like, "How do you collaborate in a team?" and "How do you communicate in a team?"

A week later, I had a phone interview with two very friendly Data Engineers. We talked quickly about my background and their jobs at Spotify. Then, I was asked many trivia questions about CS concepts, data modeling, MapReduce, Spark, probability, probabilistic data structures, machine learning, and Unix. They basically want to see the breadth of your knowledge, but also its depth in each domain. It's okay not to know a few answers here and there.

Then we switched to a shared online editor and were asked:

• A simple recursion question, plus a follow-up on how to optimize the time complexity of my solution.
• A few SQL queries, and a follow-up on what would happen if the data would not fit in memory.
• Transform one of the SQL queries into a MapReduce job.

I got a few minutes to ask questions at the end.

I was invited for the on-sites in Stockholm two weeks later. They paid for the whole trip.

There, I had two interviews in the morning, one lunch, and then two more interviews. Each interview was with two engineers and lasted one hour, with ten minutes for me to ask questions.

• Interview 1 was the "Case interview": I was given an architecture and a production problem that really happened at Spotify to solve. I could ask details to my interviewer as if he was my computer (check log files, HDFS file sizes, etc.). I was advised to be methodical, ask questions, and be clear about my thinking and the assumptions I was making. This was an extremely fun interview!
• Interview 2 was the "Programming interview": basic coding questions about API design, unit testing, and solving an ETL problem using a data processing framework (no need to know the exact API, at least the big picture).
• Interview 3 was the "System design interview": A simple question at first, but I had to ask a lot of questions to clarify the problem and make reasonable assumptions. I was advised not to be afraid to go to the whiteboard early. I was told to solve the problem for simple cases and then think about scale and how I would spread the data on a few machines. CtCi system design questions definitely helped for the general way of thinking in this kind of interview, also Jackson Gabbard's video on system design. I was also advised to read about data architecture and real-time data processing.
• Interview 4 was the "Culture fit interview" with two managers: An interesting conversation about my background and my vision about teamwork, conflict resolution, interest in my job, what I'm proud of, and times when I failed, etc. They go deep into details, so I was advised to think about these subjects and have an example for each one of them (a time when it happened during my career and how I handled it, successfully or not). I was also advised to check out Jackson Gabbard's video on behavioral interviews (he really likes his videos).

I heard back the week after and accepted the offer.

Overall, it was a smooth process and the people were very nice. It seems to me that Spotify focuses more on culture fit and collaboration than other companies I've interviewed at.

Can't give out details, but see answers above for advice.