4 ms·
With interviews, you are essentially trying to predict the job performance of a candidate when they are at your company, so why not make it as close to that as
by loftyal 8y ago
With interviews, you are essentially trying to predict the job performance of a candidate when they are at your company, so why not make it as close to that as possible?
With face-to-face tell them to bring their laptop or you can provide one, then explain a real actual hard problem you've solved or one of your colleagues has, that took roughly 2 hours, leave the room and let them try and solve it on their own accord without any pressure of people staring at them whilst coding. Come back and ask them to explain their solution and quiz them on different parts.
This has been by far the best test I've given that predicted real on job performance.
- jl2718 8y agoThis seems really obvious. Why do you think nobody else does this?
- loftyal 8y agoWhiteboarding, Programming and Comp Sci Trivia is pretty rampant...
- ggggtez 8y agoTo be a true scientist, you shouldn't necessarily do what seems obvious. For all I know, maybe whiteboarding is a better indicator of performance than 1 hour of coding on a laptop. How well does it compare? I bet that's a closely guarded secret.
- AnimalMuppet 8y agoI bet that it's unmeasured, and maybe unmeasurable.