|
|
|
|
|
by ccajas
2159 days ago
|
|
> People easily identify that white-boarding interviews suck for more than just the interviewee. And yet any time a solution is proposed that involves doing something other than putting the candidate in a room in front of a whiteboard for an hour, it's shot down. Most people take issue with the repetition of similar tests and evaluations for multiple companies. If A, B, and C were considered different stages of a technical interview, candidates are repeating A and B too often, whereas applying A and B once would suffice for several companies. It's a problem that is realized when opportunity cost is compounded when interviewing multiple companies and it's not something that a single company alone can fix. A more "Docker-ish" solution where using one of "A" to lay the foundation for many company interviews (and not just the ones expecting you to practice Leetcode) might curtail the problem of using more time and resources than necessary. But more than that, there is also a cognitive dissonance with the beliefs of how to reform a broken system. Programmers largely oppose industry-wide regulation while also expecting the vetting process to be more consistent and make more sense. This is a tough nut to crack because we seemingly want to have it both ways. |
|
...of programming (because the regulation would never be able to catch up to the state of the art), sure. Of HR practices? I don't think I've ever heard an engineering lead lament that their company's restrictive HR practices prevent them from testing candidates the way they want.
> and it's not something that a single company alone can fix
Sure they can: micro-credentials. Someone needs to be the CompTIA of tiny programming challenges, where you prove once-and-for-all (in a proctored situation) that you can solve FizzBuzz or whatever, and get that fact recorded in some database.
Hopefully, candidates could then put their MicroCredentialsCorp username on their resumes or into companies' application forms or what-have-you; and then first-stage HR automation (the same kind that matches resumes to keywords) could run through a bunch of RESTful predicate-endpoints checks like https://api.example.com/u/derefr/knows/fizzbuzz (which would return either a signed+timestamped blob of credential JSON, or a 404.)
Hopefully also, allow anyone to create a test (maybe have the end of that HTTP route just represent the username+repo of a GitHub repo where an executable test for the micro-skill lives?), such that these micro-credentials can live or die not by central supply, but rather by market demand.
(I think someone suggested doing this with a blockchain once, but there's really no need for that.)