[908B64B197]

> would struggle to recognize whether dfs or dp is the best approach to some problem within ten minutes.

But they can write a solution even if it's not the most optimal one? And they know and can tell you multiple ways of solving a problem and discuss the tradeoffs? That's good enough for me.

> the hardest problems actual engineers face are nailing down requirements and figuring out who/what to consult to unblock themselves. by the time you sit down to code a solution to a well-defined problem, most of the work is already done. to put it in joel spolsky terms, lc tests for "smart", but doesn't provide any signal for "gets stuff done".

It's hard to evaluate that in an hour, and I'm not certain there's a correct way of doing it. A college degree from a good school where students are assigned non-trivial capstone projects would probably be a good predictor.