I'm not sure you appreciate the volume of utterly incompetent interviewers and processes now out there. I have 20 years of experience, a sizable catalog of open source work available for browsing, and yet even at this level I'm subjected to the nonsense that is leetcode.
Another beef - I have yet to see an exercise which is remotely relevant to the work being done. Shuffling arrays, regurgitating interview preparation exercises, mutating objects, all-out silly exercises like "convert this number to a roman numeral" don't measure a candidate.
So where do you set the bar? This goes being annoying - leetcode exercises are pure laziness and a red flag of an inept hiring process for all but the least experienced candidates. The job market solved this problem eons ago - references, references, references. If you have none and you have no portfolio on which to measure your ability, you fall below the bar and sure, then I'll concede that some measure of basic ability is necessary. No one asks for, calls, or contacts references anymore (we do). Want to know if someone is capable? Talk to the people they've been working with. Scrutinize references for obvious softballs. If you actually care about the person you want to work with, and think you want to work with someone you respect, put a modicum of effort in.
We've shifted to using tiered collaborative "projects" that we book an hour to two hours to work through with a candidate. We give them the ability to review what we'll be working on together, in detail, before the call so that they can prepare. It's also _not_ about finishing, there is no finish line, it's about collaboration, and we tell the candidates that. We've had many candidates state this was being their ability, and see themselves out. Thats FAR more respectful of a candidate than leetcode for someone who's spent 10 years delivering microservices for a profitable business. It's also far more representative of the work they'll be asked to do:
- Being given a description of work
- Pairing or collaborating
- Stating what they do and do not know
- Researching a thing
I'd hire someone that failed an array shuffling algo bullshit leetcode challnge, but worked very well with a small group for an hour, and was able to deliver good answers and make progress on a task.
> Hackerrank, even though it annoys you, is an ENORMOUS leveling of the playing field, giving opportunities to people who didn't have the money or connections or whatever to make it into a top CS program.
I hope I've made it clear that this has nothing to do with leveling the playing field, and nothing to do with getting into CS programs. I mentioned prior that I have 20 years of experience professionally, no degree, and the last 4 people I've recommended for hire had only one prior position, were bootcamp grads (one self-taught), and are excelling in every area. They're wonderful, motivated people. What Hackerrank DOES do is shift the burden of actually caring about the people you're speaking to. If volume is an issue, the process is BROKEN.