Hacker News new | ask | show | jobs
by jlees 3732 days ago
Just because you are smart enough to compete in an Olympiad doesn't mean you can function practically on a software engineering team. Whiteboard coding helps the interviewer see how you think and function as well as "can you solve this problem"; though I agree that it's not necessarily the best way to evaluate those capabilities, it's the best tool some companies have.
2 comments
vonmoltke 3732 days ago
I don't really see that. I think whiteboard coding is one of many ways a team could function and that how a candidate performs on it is not generalizable. For instance, I have never written code, pseudo or otherwise, on a whiteboard outside of an interview in over eight years of writing code full-time. Hell, I used a whiteboard more in my four years as an EE and two years as a SysE than I have as a SysE/SwE.
link
jlees 3730 days ago
It depends how you evaluate the exercise -- for example, don't mark the candidate on correct code and syntax, but reasoning ability, communication, and problem solving. Whiteboard code interviews also have the flaw that the type of problem they tend to cover is not one you might run into in daily coding, so the way the candidate thinks and communicates is seen through a very particular lens - but my point is that given an Olympiad-level candidate who can clearly solve those kinds of problems, I think there is still value to be had from watching them solve them.

I prefer whiteboarding systems design/architectural concepts, which is definitely something I do in the course of my regular work.

link
kafkaesq 3731 days ago
Whiteboard performance (positive/negative) doesn't correlate all that well with practical performance on engineering teams, either.
link
jlees 3730 days ago
Oh, I can believe that. But the way said performance is measured -- both types -- is extremely varied. I'd summarize my viewpoint as "whiteboard interviews tell you something about how the candidate thinks and communicates, which you can use to jump into deeper explorations". Just because you know someone is smart enough to solve a problem doesn't mean it's not valuable to see how they do it.
link