It seems tech has taken the "CS first principles" side of whiteboarding interviews and applied it to ops roles, where "first principles" means "build an entire monitoring system from scratch"
Several of my SRE/ops friends have job hunted in the past year, and some of the technicals they've encountered have been contract-level amounts of work
"build Nagios from scratch"
"build Puppet from scratch"
"should have its own DSL"
"should have its own testing framework"

these are contract briefs, not interviews
I have SEEN these take home assignments and I am not exaggerating at all

experienced industry professionals are losing days to these requipmen
the brief usually reads like some internal engineer made their own bespoke system and then reverse-engineered it for an interview because "that sounds fun"

so now the candidate has to recreate this genius's design from black box descriptions
I just saw one of these technicals that explicitly stated there was no time limit and to "show us your best work"

so, let's just write off parents and people with 2 jobs from the get-go, then
honestly, if you can recreate Nagios or Puppet from scratch in an evening, why the fuck would you want to be a senior ops engineer for a branded hoodie and some RSUs
I'm still mad about this

what is the SIGNAL you are trying to get?

"understands how Puppet works" ask them how Puppet works in an interview
"can code" give them a coding challenge or have them review a PR
do you have your software engineers build an IDE from scratch in their interviews? because having an ops engineer build Puppet from scratch is the same damn thing
You can follow @alicegoldfuss.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: