It’s no secret that finding awesome talent is hard. Really hard. We’ve recently opened a position for a Senior DevOps Engineer and to qualify candidates for that position, we created a shiny new challenge. We had candidates apply and some of them provided the much appreciated feedback that our challenge was too long.
I want to find out if that’s true.
I value our candidates’ time. We want these challenges to be just long enough so candidates can provide substance, prove their chops and give us some great material for follow-up discussions with the team during the interview loop. That usually puts them at less than a handful of hours spread out over a week or so.
The challenge’s history
When we put this DevOps challenge together, we had a very specific quality bar in mind: world class. We come from a world where we git push to production multiple times a day code that’s gone through continuous integration tests on every commit. Merges into our services’ master branch get auto deployed to our production-like environment that undergoes continuous testing through not only armies of testers in the cloud but also our partners that poke and prod at our products and API. We have Flowbot help out with deploys and organize ping-pong matches. And we’re just getting started, no one is on this full time. And yes, we use Heroku to host many of our services because we want to focus on building products for users. But we’ve also invested in Docker and AWS to deploy, isolate and scale the variety of adapters we’ve built to support the world’s largest sellable bus inventory. As we scale our stack, we want to bring more and more of the fine-grained control those technologies offer over our infrastructure back into our hands so we can offer a better, faster experience to our users and partners worldwide (like our recently added partner Voyages SNCF).
To get to where we are today, we stood on the shoulder of giants (Heroku) and learned. A lot. In fact, I fondly remember Dustin telling me he learned Go, contributed to Docker Machine core, learned Terraform and he and Valerian spun up our baseline infrastructure. And all this in less than a week.
So I have no doubt our team can step up to the challenges DevOps face and deliver. I just don’t want them to. At least not all the time. We’ve got other fish to fry too, and doing infrastructure right requires an unrelenting focus. There’s a lot at stake, and we respect that. Hence the job post.
Which brings us back to the challenge. And Donald (Rumsfeld, not Trump).
There are known knowns. These are things we know that we know. There are known unknowns. That is to say, there are things that we know we don’t know. But there are also unknown unknowns. There are things we don’t know we don’t know.
– Donald Rumsfeld
Busbud is looking for an expert in DevOps. Someone to raise the bar for the team, someone that will look at that challenge, crack their knuckles, roll up their sleeves and bang out a solution in hours. But maybe this is a story of unknown unknowns. Maybe the right solution isn’t possible in a matter of hours, but is more the stuff of weeks or months. That’s where you come in. You tell me if our challenge is nuts.
How you can help
Here’s how you can help:
- Take a look at the challenge
- Estimate how long it would take to complete the challenge to show your level of competency (and maybe a separate estimate if you believe the production version would be materially different)
- Let me know your estimates via comments below, this form (if you’re into that) or Twitter #DevOpsChallengeChallenge
Some might say this is a clever marketing ploy to get more eyeballs on the job post. Perhaps. But I can tell you I am legitimately interested in understanding if our challenge is too long for the bar we set, and this is my best bet at coming up with data to answer that question.
So again, thanks for your help.
Co-Founder and CTO
PS If you use the form, and we have enough responses to make it interesting, I’ll write a follow-up post to share the results.