Hacker News Reader

On my first Amazon interview, I'm pretty sure I bombed because I brute forced the algorithm question.

In the real world, when we have an issue on production, I fix it first, then find a solution. Meaning in that first hour customers are complaining, I'm not gonna try to create a fancy algorithm that can handle all cases. I'll make it work with nested loops if need be. Then I can sit down and figure out a permanent and elegant solution.

View on HN

FWIW, that's typically the structure that Amazon is looking for in an interview (though, some interviewers are better/worse than others)

- start with the super inefficient solution, solve it pretty quickly

- point out performance problems or edge cases that would need to be solved for a larger scale solution

- go implement those optimizations

Yup, when I was interviewing for Google I was expecting basically the same. I would like you to work though the problem and come up with something simple, bonus points if you noted places that could be buggy, slow or anything else. Then we would look at it and discuss what could be improved and if we had time maybe even make those improvements.

Reply on news.ycombinator.com