The first phase was solving some problems on HackerRank. I wasn't familiar with HackerRank before, so I prepared by solving some problems on it a few days before the test. On the test itself, the questions were not particularly hard in and of themselves, but the time was very limited and didn't give me much chance to sit down and really think about what I'm doing. I'm not particularly good at relying on "programming reflexes" rather than thinking deeply about the problem, and ran out of time on the last question, but I got through anyway.

The second phase was on-line interview with two people from G-Research, through the HackerRank's pair-programming platform. They asked me to describe interesting projects I worked on, what testing strategies I used etc. After that, we started writing code in C# and at one point they asked me about implementing some cache eviction strategies. Implementing caches not being a part of my day-to-day job, I fumbled at first, but asked additional clarifications and got it right in the end. In all honesty, I did implement some caching code few years ago, but that was a VERY different case (with multi-threading and completely different eviction strategy), so I needed some time to get my bearings right. However, my interviewers thought I was not fast enough, so they decided to end the interview.

I think that's odd. Apparently they want engineers who win "interview lottery" (are familiar with the question a-priori), instead of somebody who can cope with an unfamiliar problem.

Also, they didn't ask me anything about libraries, programming languages or databases or any of the other "technologies" I mentioned on the CV. Not even "what is a database index" type of question, even though they indicated their stack relies on SQL Server.

Overall, I wouldn't call the experience "negative", it was just odd.

Implement "most recently used" cache eviction strategy (in the context of a memoized function).