On-the-job observations are very subvertable though. They happened a few times in my high school in Soviet Union. At least twice the teacher would prep "good" students a few days in advance: I will ask you this question, and this is the answer you are supposed to give, make sure to memorize it well. So the class is engaged, gives good answers to hard questions, except this is all a Potemkin village deal!
Any human can manipulate their assessment. However, I think a skilled and trainer observer has a better chance of detecting that than any standardized blind mechanism.
Practically speaking your interview would require applicants to apply claimed knowledge to novel context.
A simple method would be to identify interesting problems from recent PRs and ask them to walk you through their approach to discovery and solution. It's a problem they should be familiar with, but in a new shape and with different labels. Let's see what they come up with.
