A 45-minute phone screen interview with a simple question.

On-site interview:

Five 45-minute interviews, plus the lunch hour. At other companies, there are typically four interviews plus an informal lunch interview by the hiring manager. Google's approach came as a shock; the person taking me to lunch wasn't even a manager, so they couldn't answer many of my questions.

None of my questions were answered to my satisfaction because five out of six people I met did not work in an area remotely related to the team I would be joining. The common answer was, "it depends on the team."

In general, the interview was tailored for someone fresh out of college (really smart but with little experience), even though the recruiter told me I was being interviewed for a Level 4 or 5 engineer position.

3 Algorithmic Questions

I say algorithmic because the coding questions are one-size-fits-all. They are designed for Python programmers. If you write the code in C, it doesn't give the interviewer any indication about your deeper understanding of C programming (such as defensive programming, memory management, dangling pointers, cache efficiency, etc.).

In most interviews, I couldn't get straight answers to my clarification questions. It was always, "however you like it." An important part of system design is to understand the constraints of the system and plan accordingly. When I am not given any constraints, this means I can write whatever I want. They failed to measure this skill.

The algorithmic coding questions were not structured properly (no multi-part questions). It seems they have no way of making them easier or more difficult. When I finished the problem with the optimal solution, the interviewer asked some weird questions instead of giving me a challenge.

The system design question was way out of my league. I am interviewing for an embedded engineering position, yet I was asked to design a distributed network-based system. None of that applies to my domain, and none of my experience helps me answer the question.

The embedded question was the only one that measured my experience. The questions were not really deep but satisfactory.