As VP of Engineering, hiring gets an important share of my attention and agenda. Staffing makes-or-breaks a team, thus why I pay so much attention to how we hire and who we hire.

Lately, after negotiating fiercely (I didn’t have to) with the CEO for hiring 4 extra software engineers, I though it was a good time to empower my team to revisit our hiring process. Of course, some guidelines remained: keep it humane, empathise with the time we steal from each candidate and the stress we cause them. Above all, let’s answer the question “would we work well together?” which is a 2-way street, thus need to be embodied by the hiring process.

Just in time, Barry Cranford from RecWorks and the London CTOs community announced a roundtable betweem senior engineering leaders on the topic of hiring with Adam Martin, CTO and author of “Hire the first 100”.

Quotes

21:47 “What are “top candidates? They are the people who will have the greater impact in their role.”

My key takeaways

Feeding your hiring pipeline

  • How to find more good people? The recruiter should not only show interest for the candidate they are speaking to, they should also ask the candidate who they talk to when they get stuck. For example:
    • Recruiter: “who is your go-to person when you struggle on a technical problem? Candidate: “Oh, it’s John Toddler.” Recruiter: “would you happen to have his contact details?”
    • Of course, this is not to extract the best-of-the-bests but it helps multiply the number of good candidates.
    • Similarly, identifying the authors of relevant comments on technically strong blog posts represents a good pool of candidates for your hiring pipeline.
    • 39:18 Yet another similar approach is to check the GitAwards website to see how people are ranked by their community based on their github contributions.

Screening and take-home exercises

  • 5:59 A participant said they hire 20 engineers a month and don’t observe that take-home tests correlate with performance afterward. It is useful only when it supports a conversation thereafter.
  • Take-home tests become useful when the job requirements have strict constraints like strong skills in C++, Asm, or Python.

Avoiding bias

  • 17:34 For one participant, the technical skills matters less than the ability to communicate clearly due to the essential complexity of their business domain.
  • 19:32 Suggestion to evaluate each candidate with less bias.
    • Define 20 questions that will be asked to every candidate.
    • Grade every questions: first question is worth 3 points, second is worth 5 points, etc.
    • Score all candidate’s answers
    • Compute the final score of each candidate.
  • How do we remove bias from interviews? “Cultural fit” tends to favor same origin, same studies, same pay grade, etc. rather than talent.

Quality of your hiring process

  • Do you measure the quality of your hiring process? beyond simply asking the candidate for feedback?
  • Complex, large, vague, questions like the Google question are useful to keep your senior engineers engaged with the general hiring process. Otherwise the answers tend to be so shallow or plain wrong that they aren’t really useful to assess the candidate’s hard skills.
  • How to train interviewers?
    • 48:48 The Shadowing process is the core of NHS (National Health Service) teaching. It is brutally efficient to train surgeons. It can be summarised as “see one, run one, teach one”.

References