3 ms·
Good values aside, this is deceivingly brilliant as a hiring technique. Hire the undervalued "hungry" people who the Googles/Facebooks of the world won't even
by ajaymehta 14y ago
Good values aside, this is deceivingly brilliant as a hiring technique.
Hire the undervalued "hungry" people who the Googles/Facebooks of the world won't even give an interview, and they'll 1) be easier/cheaper to convince and 2) work super hard to prove themselves (especially if it's their first job out of school). If the startup does well, it's a win-win.
- thedufer 14y agoBut without the filter of "top schools", how many candidates do you have to sift through? People use that criteria because its easy and means something (although maybe not much), not because they think it's perfect. You can only afford to spend so much time recruiting talent.
- ajaymehta 14y agoYou can still sift through resumés, but don't filter by school. Filter for people who have "created something" or who have a thoughtful online portfolio.
- wnight 14y agoCode or GTFO. I ask trivial coding questions (fizz buzz level) but take them further. First write that, then change it a bit, then more, then make it customizable, etc. I also like the idea of saying 'mention (some shibboleth) in your cover letter' to prove people read the add. If I was writing an ad I'd ask someone, maybe, to provide their favorite code snippet and describe why. Anything would be okay, as long as someone has an articulate reason. I get disappointed in an interview if the company don't want to see code. Someone in this thread discussed how school doesn't correlate to the ability to ship code; I'm good at having an executable answer to a problem in minutes. If they don't want to see what I can do it usually seems like it's because they don't know how to judge it, and thus aren't who I'd like to work with.
- thedufer 14y ago> I ask trivial coding questions (fizz buzz level) but take them further. First write that, then change it a bit, then more, then make it customizable, etc. This doesn't even come close to scaling when your applicant pool is 100 (or more) times the size of your interviewer pool. You have to have some quick filtering method, or you'll never get anything done.