The steps for me were:

1. Online pre-screening with a recruiter (~1h).
2. (If the recruiter is happy) Online screening (~1h) with two engineers. The first half involves discussing two or three relatively simple algorithmic problems, exploring multiple solutions, optimizations, and the complexity of those solutions. The second half consists of trivia questions, quite a few of them, some very concise, others more open-ended where you can ask questions for clarification.
3. (If the engineers are happy) On-site interviews - four interviews (~1h each) in one day. Two interviews were scheduled before lunch, followed by lunch with random employees from different teams, and then two more interviews. Each interview was conducted by two people:
   • One with hiring managers, focusing more on soft skills and cultural fit.
   • One case-study interview (incident troubleshooting and problem mitigation).
   • One system design interview (with an angle on data processing, as I was interviewing for a Data Engineer position).
   • One coding (on the whiteboard) interview with an additional angle on testing approaches.

Overall, it was a pleasant experience, although the on-site interviews were quite exhausting.

The recruiter stayed with me throughout all steps, checking in, asking for feedback, and forwarding very good and concrete feedback from each interview.

The process was quite long due to the Christmas and New Year period, affecting the availability of interviewers as well as my own personal availability.

I received feedback at most a week after an interview regarding whether there would be a next step.

Given a firehose that shoots real-time events about songs being played, how would you calculate the top N tracks played in the past day, and keep those top N updated through time?