While that works for entry level jobs where direct algorithmic knowledge is the most important, it's hard to test if someone can lead a team of 6 to finish a quarter of a million dollar project on time and on schedule through a 2 hour whiteboard conversation.
Algorithmic tests aren't great for any level. You want to ask someone to solve practical technical tasks that are relevant to the job they are applying for. For an entry level job this may be a programming task (build this one screen). For a more senior position you'd want to ask about things like architecture and practical trade-offs. You could use a whiteboard for this: I don't think it's whiteboards people object to. It's the irrelevant algorithmic questions.
Very true. I didn't use the right term there for what an entry level dev does. But I still think that, even if a senior tech can do a decent job elucidating a architecture, that's not proof they can lead, which is similarly important, and the only way to prove that you have the ability is to look at past accomplishments.
> the only way to prove that you have the ability is to look at past accomplishments
I'd argue that past accomplishments don't prove anything either. Or rather, it's impossible to tell whether they were actually responsible for those accomplishments. It's not just architecture you want to ask about. It's the practicalities of choosing libraries, making engineering trade-offs, running a team (if that's part of the job they're applying for). The trick is to ask them questions that they'd only know the answer to if they've actually been there and done that ("What want wrong" and "what problems did you face" are good ones). And to follow up with asking them why they'd make that choice.
