[maxander]

Thinking like an engineer: to hire good engineers, you need A) good engineers to apply for the position, and B) to tell which of your applicants are the good engineers. Those two conditions are necessary and sufficient.

As other commenters have noted, A boils down to “offer lots of money.” Some really interesting fields may be able to offer less out of “cool factor,” but odds are you’re not hiring for one of those fields- all your potential hires who claim to think CRUD apps are fun and intellectually stimulating are just lying. ....So A is simple, if you’re honest with yourself and bite the financial bullet.

B is a problem that is apparently beyond present-say technology, and quite possibly not possible to achieve at all. There are code-testing sites that go beyond the standard CS interview, but how can a several-hour exam test someone’s ability to manage a codebase over months or years? In particular your company’s codebase, which will have many important idiosyncrasies and be a very different environment than any project the hire has worked on in the past. ...We’ll likely have flying cars before we figure out a halfway decent solution for B.