The interview process was simple: an initial phone screening, a Google+ Hangout interview with a coding exercise, and an onsite interview that lasted 3 hours with 3 separate interviewers.

• The initial interview went well, with no problems.
• The second interview is where things started becoming awkward. I began with a senior engineer who asked a few standard questions, as well as more complex topics like garbage collection and distributed systems. I was able to answer most of them, though I was not applying for any tuning or infrastructure roles. The coding exercise was strange. He asked a question about how to find the maximum value in an array, an easy O(n) problem that I implemented perfectly. I mentioned that finding the maximum could be done using a fold operation. As soon as I said that, he seemed to become confrontational, immediately asking me to implement fold, which I did. Then, with 10 minutes left in the interview, he asked me to complete a maximal subsequence problem in Java, a language with which I am not most familiar, as I am a Scala programmer (the role they were hiring for). I provided the basic algorithm, but because we don't see dynamic programming often in Scala, it was a tough sell, not to mention an impractical amount of time to ask such a question.
• The onsite interview reflected more of the same. I came prepared for more dynamic programming questions and nailed the first and last interviews. However, the second interview, focused on constructing a distributed system, was not my best. To add to that, the interviewer used terminology that I don't think he understood (for instance, asking to construct a real-time event feed, but then describing a solution that was more of a micro-batch update feed). His explanations were confusing to someone who understood the terms.

The takeaway from all of this is that the job listings are not quite accurate. The interviewers are very naive when it comes to interviewing, particularly regarding the time allotted for answers and the content of the questions. Since when are vector clocks and in-depth knowledge of garbage collection algorithms considered fundamental computer science concepts, especially for a functional programming role?

Construct a distributed system that handles real-time event processing.