2.2 Interview pipeline
The interview pipeline for ML roles evolved out of the tried-and-true software engineering interview pipeline and includes the same components that one can see in a traditional technical interviewing process.
There are many people involved in the interviewing process. For hiring managers, it’s crucial to assign each interviewer a set of skills to evaluate, so that different interviewers ask different questions and that collectively, they get a holistic picture of where you’re at with all the skills they care about. One interviewer might ask you about theories, a couple about coding, one about ML systems design. Ideally, the interviewer tells you the skills they want to focus on so that you can tailor your answers to highlight those skills. If they don’t, ask.
Recruiters are often encouraged to share with candidates the names of their interviewers. If they don’t, ask. You can look up your interviewers beforehand to learn what they do. It’ll give you a sense of not only what you’ll be doing if you join the company, but also your interviewers’ areas of interest.
You should also ask about the team you’re being considered for. At most companies, you interview for a specific team. However, if you apply to companies such as Google and Facebook, you’re matched with a team after you’ve passed. This means that it’s possible to pass their interview process without getting an offer if no team takes you, though it’s rare40.
The following list of interview components is long and intimidating, but companies usually use only a subset of them. The number of interviews in each component also varies. Companies might skip any step if they’re confident about your ability. Strong candidates might even be invited directly to onsites without all the previous steps. On the flip side, candidates that need to travel for onsites might be vetted more rigorously beforehand.
The entire process can be long and tiring, but it’s long and tiring for every candidate. You don’t have to answer all questions flawlessly. You only need to do better than other candidates. No company expects you to know everything. The field is moving so fast it’s unrealistic to expect any candidate to know all the latest papers and techniques.
40: How many people failed at the team matching phase when interviewing with Google? (Quora)