3 ms·

You're basically describing HackerRank, and I've worked with some pretty terrible engineers who were very proud of their HackerRank credentials. Write such a t

by macksd 4y ago

You're basically describing HackerRank, and I've worked with some pretty terrible engineers who were very proud of their HackerRank credentials. Write such a test. Give it to all the engineers whose work you know well, and see if it can accurately predict what they brought to your team. I would be very surprised if you could both spot good talent and dead weights reliably.

bambax 4y ago
It doesn't have to be perfect, it just needs to be better than the tech part of the current interview system. I very much doubt that whiteboard interviews can accurately predict what people bring to a team; they're not even designed to assess that. This system wouldn't replace the personality interviews, etc., but only measure technical abilities.
- macksd 4y ago
  Is there much actual whiteboarding going on in tech interviews right now? It's been years since I did anything but remote coding interviews (where you have a fairly realistic editor and REPL environment) and remote behavioral interviews. I liked doing whiteboards where someone would walk us through a former project and architectural decisions, etc. with diagrams. I kinda miss that, actually. But I haven't seen literal whiteboard coding since long before COVID even.
rkangel 4y ago
We had some fairly standard interview questions that I wrote - problem solving type questions asking someone to sketch out a pseudocode answer. One of the smarter things I did was 'interview' several of our existing employees to get an idea of where people would fall on the scale. It was a great help when we then started giving them to candidates, and prevented us hiring some people because no, the good people we had really could solve these problems with only a bit of nudging.
- macksd 4y ago
  This is more or less what we do on my current team. We write out a doc explaining all the common solutions, and what a good candidate should consider and be able to answer, and have a rubric, and we ensure people don't judge candidates until they've shadowed enough that we should have a pretty consistent scale and have made a reasonable effort at being objective. The downside is some people have barely scraped by in the interview and turned out to be absolute rockstars and day 1, and some people have aced the process and driven everyone else mad with incompetence. So it could be improved. And maybe that's just on us not writing the perfect questions yet, but a bigger problem is that once we've used a question enough someone will post it online and suddenly a suspicious number of candidates are able to show up and rattle off all 3 optimal solutions in 15 minutes. Weird! I suspect a standardized test approach to software engineers would get this problem even worse.

by macksd 4y ago

bambax 4y ago

It doesn't have to be perfect, it just needs to be better than the tech part of the current interview system. I very much doubt that whiteboard interviews can accurately predict what people bring to a team; they're not even designed to assess that. This system wouldn't replace the personality interviews, etc., but only measure technical abilities.

macksd 4y ago
Is there much actual whiteboarding going on in tech interviews right now? It's been years since I did anything but remote coding interviews (where you have a fairly realistic editor and REPL environment) and remote behavioral interviews. I liked doing whiteboards where someone would walk us through a former project and architectural decisions, etc. with diagrams. I kinda miss that, actually. But I haven't seen literal whiteboard coding since long before COVID even.

rkangel 4y ago

We had some fairly standard interview questions that I wrote - problem solving type questions asking someone to sketch out a pseudocode answer. One of the smarter things I did was 'interview' several of our existing employees to get an idea of where people would fall on the scale. It was a great help when we then started giving them to candidates, and prevented us hiring some people because no, the good people we had really could solve these problems with only a bit of nudging.

macksd 4y ago
This is more or less what we do on my current team. We write out a doc explaining all the common solutions, and what a good candidate should consider and be able to answer, and have a rubric, and we ensure people don't judge candidates until they've shadowed enough that we should have a pretty consistent scale and have made a reasonable effort at being objective. The downside is some people have barely scraped by in the interview and turned out to be absolute rockstars and day 1, and some people have aced the process and driven everyone else mad with incompetence. So it could be improved. And maybe that's just on us not writing the perfect questions yet, but a bigger problem is that once we've used a question enough someone will post it online and suddenly a suspicious number of candidates are able to show up and rattle off all 3 optimal solutions in 15 minutes. Weird! I suspect a standardized test approach to software engineers would get this problem even worse.