First, there was the algorithmic section. Two problems. The first was very simple. The second was about topologically sorting a graph. The interviewer didn't know the classic Kahn algorithm for solving this problem. When I tried to solve it this way, he said it was wrong, that there was no such solution. He spent about 15 minutes convincing me that such a solution doesn't exist, that it can't be applied, and that it's incorrect. I only wanted a DFS solution. You not only need to solve the algorithmic problem, but also solve it using a non-optimal method that the interviewer knows.