Your process seems equally flawed. Many excellent candidates have no open-source code, and if they show you any closed-source code that's an immediate signal against hiring them.
So now you've put candidates through a silly "live coding challenge" – offering an advantage to the sort of person who is chatty and communicative with strangers in a stressful environment and the expense of people who may be better engineers but slower to warm up.
I really wish folks would be way less dogmatic about this stuff. Every single time I see someone say "take-home exercises are bad", their proposed alternative is also critically flawed.
It might be flawed in terms of filtering out the type of person that does not do well in live pairing scenarios.
But I disagree it's _equally_ flawed. The time investment from a candidate is significantly smaller at the cost of more time investment from the employer's side (aka my side). Personally I consider this a reasonable trade-off.
So now you've put candidates through a silly "live coding challenge" – offering an advantage to the sort of person who is chatty and communicative with strangers in a stressful environment and the expense of people who may be better engineers but slower to warm up.
I really wish folks would be way less dogmatic about this stuff. Every single time I see someone say "take-home exercises are bad", their proposed alternative is also critically flawed.