Interviews#
Interviewing is hard. It’s also important to do well.
A good/bad match can be transformative/terrible for both the company and the candidate. This article describes my current approach. It may be useful for others who are conducting interviews, or for people who plan to interview with me (transparency is good).
My interviews are typically in 45 minute calendar slots. After saying “hi” and being friendly for a minute or two I lead the conversation through the following stages:
I talk about the company and what we’re looking for#
This typically takes five minutes
First, I give information about the company, the teams within the company and where we think the candidate might fit, and what kinds of candidates we think work well with us.
In our case I say things like the following …
We’re a commercial company based around the open source Dask project
Engineering-wise we do two things
Work on Dask and push it forward
Build a cloud SaaS deployment platform
Given your background, we think that you’re a good fit for one or the other, but our most impactful folks end up being able to cross both.
Historically we’ve found that we work well with senior engineers who are professional (can deliver things on time, even in complex situations), know open source culture, but also care about commercialization (we’re a for-profit company and culturally lean into that). This is what we’re looking for.
On the last bit I like to be explicit. People need to opt-in to both our strengths and our constraints:
Seniority: we’re bad at management and don’t mentor people well when they’re early in their careers
Professionalism: there are so many interesting problems around us that it’s easy to get distracted (we all suffer from this). We prefer important work to interesting work, prioritize intentionally, and feel good about setting commitments and working to timelines. This is typically hard for academics.
Commercial OSS: we believe in the power of open source communities. We also understand their limitations. We are excited about commercial impact and and using that energy to both make money and have social impact.
There are very good people who don’t fit that culture. We like to be explicit to avoid mismatches. Alignment is good for anyone.
Learn more about the candidate#
This typically takes twenty minutes
Next, it’s time to figure out if the candidate knows what they’re talking about. There are many ways to do this, none of which are particularly good.
Personally, I like to ask them what they know about us and if they have relevant experience. I then like to dig into that experience and test different situations to probe a bit.
Interviewer: “Have you used Dask or anything like it before?”
Candidate: “I haven’t, but I did use Spark at my last job”
Interviewer: “Oh cool, what did you use it for?”
Candidate: “We did a lot of machine learning and AI”
Interviewer: “Oh interesting, what kinds of models were you training?”
Get to probe their understanding of technical work
Candidate: “We tried using XGBoost, but really most of it was logistic regression”
Interviewer: “I’m curious, I don’t know much about this field, but my understanding is that it’s commonly better to just sample and train locally. Why use distributed computing at all?”
See how they react to a challenge. See their ability to critically analyze their own directions.
…
Interviewer: “Why were you training these models?”
Probe their understanding of how their work affects the rest of the business
…
Candidate asks questions of us#
My usual line here is “I have a couple more questions to ask, but first, do you have any questions for us?”
What they ask here is telling. Sometimes people are ask generic questions. These look like the following:
What are your biggest challenges?
How is company culture?
Where do you see yourself in five years?
More informed candidates usually have more specific questions for us:
How are you balancing between the cloud platform and OSS work? Are those teams siloed? Is there space to work on both?
Tech companies are in a bad way these days. How secure is my job here?
I’m really interested in application X. Is that something that there would be space to work on while here?
Often I’m not judging them at this point. I’m probably trying to sell them. It’s good and healthy to have the tables turned. Interviewing is bi-directional in any good relationship.
Logistics questions#
At the end I keep about five minutes to ask the following:
What’s your risk tolerance?
Mostly I’m trying to scare them:
We’re an early stage startup. We’re not shy about changing direction. We have a history of firing people. You seem great and we’re picky about who we bring on, but we like to be upfront about the possibility of separation.
Typically people in this space can get a new job within our typical severance period (one month historically) but are you comfortale with this sort of dynamic?
What are your payment expectations?
This is an awkward conversation to bring up. My usual phrasing is something like the following:
You can totally pass on this next question, but roughly what your are you looking to make here salary-wise.
The reason I ask is that we’ve gone through interview cycles with candidates only to find, at the end, that they’re expecting $300k+, and we’re nowhere near that. I just want to make sure that we’re somewhere in the same neighborhood.
I then wait a bit, and typically the candidate either …
Just tells me what they’re making now
Gives a general range for what they’d expect someone in this role could make
Does a hard-pass (although this is surprisingly rare)
Most often there is also a conversation at this point about equity, vacation expectations, other benefits, which is good to have. I’m also open to sharing a broad distribution of what our engineers make.
Possible start date
Some folks have long notice periods (looking at you Europeans) others just got let go and want to start next week.
Next steps#
If I’m leading the process I’ll maybe mention a few folks that I’m going to want them to meet. If I’m not leading the process I say “I’m going to share my notes with the team and the person who’s leading this process should get back to you shortly after your last interview”.
Timing#
Speed is key. Especially in tech. Especially in startups. Sometimes we’re great at this. Sometimes we’re terrible. It’s always something to work on. Some tips:
Early-reachout once they apply, if only to acknowledge the application
Panel interviews with lots of folks at once
Pre-schedule a meeting with the interviewing team for right after their last interview