Yandex

One-hour video interview, where I had to solve a medium-level algo problem. I'd solved it, yet there were bugs in my code, so I spent another half of the interview trying to find them. Finally, I'd found them all, but time came to the end.

The interview was purely algorithmic. The interviewee knew them worse than I did. It's really strange that the company puts people like that to interview interns. I really suggest going straight for a junior job and not getting stuck as an intern.

Improving existing algorithms and creating new ones involves several key elements: * Machine learning * Working with very large amounts of data * Writing algorithms to work with them An example of this is image search optimization in Yandex.i