Startups: Learning from 3 years of technical interviews @ Improbable
As a first foray into the blogging world, technical interviewing isn’t quite the sexiest of topics, but given how vitriolic we all can get, I wanted to outline my thoughts and provide examples so others don’t fall into some easily remediable traps. This is by no means an exact science easily replicated to all needs, but rather my take on a divisive topic that can make all difference at the earliest stages of a company.
Over the course of our 3 year existence we have interviewed hundreds of high-calibre candidates, hired over 50 engineers and learned some valuable (harsh) lessons.
I think that whether you are starting on your path to hiring or looking to iterate on a current process, you must first look at the why. Why are we hiring? What specific business need does this role fulfil? Introspection at a time of growth or change can both inform an interview process but also be cathartic when understanding performance of the organisation as a whole. Once there is consensus on the business need driving the hire, try to understand how you will efficiently test the skills that you are lacking. Spending some extra time to really understand what is driving this hire and the roadmap for the work this employee will do, can be a valuable exercise in both self-assessment and forward planning. Furthermore you will be able to create clear expectations at both interview and within the role, tied to competencies that will allow the candidate to flourish.
With a clear sense of expectation, the other facet of a well-designed process is the notion of relevance. In my experience, candidates tend to enjoy interviews that reflect the type of work and projects they would be facing if they joined the company. We have found that we get the best data on a candidate when the interview is an approximation of a task they are likely to face if they join. In reality, especially startup life is not some perpetual holiday discussing complexity and binary trees; it’s stressful firefighting and hacking some shit. Test that…
And test yourself. Iterate on your interviews and measure, then iterate and measure again. Is your process giving you strong signals about the best people in your business? How are you comparing between candidates and processes?
The holy grail of interviewing is being able to identify future superstars very early on. If you take the time to get the process right the first time, it could make all the difference.
Disclaimer: Not an engineer, never will be. Just trying to catch them all. Work at Improbable, London.