After a month of interviews with companies as big as several thousand to a 20 person startup, I’ve encountered some very distinct interview processes. I wrote a study guide which was useful for people looking for new jobs, but hopefully this second part will be helpful to the people who actually like their companies and want more people to join.
Following the model that good managers use for feedback, I’m going to publicly praise specific companies and shame anonymously. Importantly, the shame section mostly includes hypotheticals; I screened out many companies before I applied, thanks to my incredibly privileged network of Berkeley alumni and Facebook diaspora.
Name and Praise:
Notion had a unique product manager interview, where I sat down with the CEO with a list of potential features. I would ask questions about the target audience, estimate engineering scope, etc, and then assign priorities to the features. I think this is fantastic signal to glean from a candidate, as you learn not only what they value, but also how they communicate and question projects. From a candidate’s perspective, it also signals that engineers truly do have a say in the product roadmap, with an entire hour of the interview process devoted to checking how they would contribute.
Robinhood’s technical interview involved a problem that could potentially be solved with multiple threads/coroutines but would more elegantly be handled with a single thread. I enjoyed the deeper dive into threading since not only is asynchronicity a difficult problem, but one that shows up fairly often day-to-day. This was the single most difficult problem I had in all my onsites. From a candidate perspective, this definitely solidified Robinhood’s strong technical reputation when I was considering offers.
Plaid’s take home project was a toy version of an app they were potentially building. Even if the company acquires similar coding signal from any take-home project, it is a magnitude nicer for a candidate to work on a relevant project. Since most take-home projects are 4–6 hours, that’s about half a day of your candidate potentially asking “Why am I making a Yelp clone?” instead of “Cool, I can see myself working on this at <company name>.”
Atrium went above and beyond with follow-up meetings. Candidates want as much information as possible to make a decision, and Nathalie, the recruiter, signed me up(without me having to ask) for talks with managers, executives, and even a Q&A with Andrew Chen, a board member and seed investor. She also set up a happy hour to let me meet the whole team in a casual context. Not only were these meetings good additional signal, Nathalie being a stellar recruiter was great signal unto herself. Good recruiting combined with a good company means good potential team growth, since candidates will never be lost for the sadly common reason of poor recruiting (raise your hand if you’ve ever been ghosted by a recruiter).
Anonymous/Hypothetical Shame:
Check your loaner laptops. No company got this completely right and in all fairness, it is incredibly difficult to get right. Android Studio updates every few months and some versions are straight-up incompatible with the same project code. Extend this to all platforms, and you have a full-time job maintaining perfect interview-readiness on laptops. Things companies can do to mitigate this is to ask for candidate preferences for software on loaners before the onsite (potentially in the same email where they describe the onsite). Also ensure there are both Magsafe and USB-C chargers in the interview room, even for candidates who bring in their own laptops.
Everyone should be on the same page. I ask the same question about where the company is going to everyone from IC’s to executives. When the answers are different, even if they’re all positive answers, I worry about how the company will execute its goals when direction is so fragmented. Of course, while focus and scope of the goal will be different amongst employees, there should be an overarching theme. If I sensed dissonance, that was a huge red flag.
Teams that don’t joke. This one is highly subjective, but I am pretty friendly and silly with everyone, including coworkers. If the team is there to just put in their hours and go home, I’m not going to be happy and productive there. Honestly, I loved that the Plaid take-home project was littered with puns like “Plaidillac” and “Plaidypus” and while this wasn’t the DECIDING FACTOR, it was icing on all the other positive traits.
Desks being full after 6pm. Also subjective, but I’m done with daily long grinds. People being in the office after 6 is fine especially if there’s dinner and the people who are there are just hanging out, but people actively working after 6 is a yellow flag. This is only a yellow flag, because there are several legit reasons besides bad work/life balance. Some engineers work best at bizarre night hours or want to wait out traffic and the company is flexible about hours. Maybe there’s a huge one-time deadline like a conference or a blockchain split. The first reason actually becomes positive signal and the latter remains a yellow flag. So always ask the recruiter or the manager if you see this.
General tips
Break away from the FAANG style onsites. Especially for smaller companies that don’t immediately have their questions leaked on Leetcode, take time to construct good interview questions, and potentially new formats. As mentioned earlier, I loved Notion’s project prioritizing interview, but many other companies are also starting to test debugging and code review, all things that are imho more useful signal than “Can you throw breadth-first search at this problem?”
Take-home project, add-on feature onsite. This is the format that I believe most authentically represents day-to-day work and therefore gives the most accurate signal of a person’s technical ability. While the take-home project is long and frankly, pretty much the same across companies, I liked that it was no-pressure. Even when I was working, I preferred devoting an evening to a take-home over scrambling to find a conference room for a phone screen in the middle of the day.
The second part, adding a feature onsite, is a vital portion though, since it checks for three things:
- that you weren’t cheating with the project
- that you can handle the day-to-day work of working with an existing codebase
- that you architected that existing codebase in a way to make it easy to add new features (even if you have no idea what those new features could be)
Controversially, I would like it if there was a timed submission box that was only open until 4–6 hours after viewing the project spec. It could even secretly be open until the next day, but I think the relief from setting the clear expectation of “Hey we don’t expect a complete, polished project” outweighs the anxiety of the timer. Otherwise, it becomes very easy to fall into a trap of “Just one more feature” and then you end up with angry candidates on Glassdoor complaining that they spent 12 hours on a project and didn’t get a callback.
Conclusion:
This is the process from which your company is built and should be treated as a fundamental tenet of your company; failing to recognize that is terrible signal to candidates. Be extremely conscious of what signal is portrayed with each of your interviews, as even candidates who don’t join your team will influence future candidates. I am a terribly lazy person and if even one engineer I respect tells me they had an awful time interviewing with a company, I’ll take that excuse to not apply. On the other hand, I have been rejected from companies that I had fantastic recruiting experiences with and will still gladly recommend them to people. In that vein, I suggest you check out the named companies above if you’re looking.
Let me know great and horrific interviews you’ve had!
Sidenote:
Climbing. Hilariously I had seen at least 3 people on my interview panels at separate companies at the rock climbing gym. They recognized me.
I am not sure if climbing can or should be exploited in the recruiting process, but I personally wouldn’t say no to some sweet company swag chalk bags (*cough* plaidypus chalk bags *cough*).
📝 Read this story later in Journal.
👩💻 Wake up every Sunday morning to the week’s most noteworthy stories in Tech waiting in your inbox. Read the Noteworthy in Tech newsletter.
