“Do you prefer to work independently or on a team?” has no safe middle answer. “Both” is the easy answer. It tells the interviewer nothing. The question is about two risks. One engineer goes quiet for a week and surfaces with the wrong thing. Another can’t start without a meeting. So name your honest default first. Then name the two moments you switch out of it. The first is before you start, when you show others the plan. The second is when you’re stuck, after a limit you set in advance. Back both with one project, including a time your default cost something. An interviewer can check that answer. It also shows you know your own failure mode. Below, an Android engineer builds it from one download bug she sat on too long. The last section covers the question to ask back.
Saying both to work independently or on a team gives the interviewer nothing to write
Interviewers write notes during the round. Later, someone reads those notes and decides. “Works well alone and in groups” is not a note anyone can use. It fits anyone. Nobody can check it.
Picking one side fails in a different way. “I prefer to work alone” sounds like you’ll avoid code review. “I prefer a team” can sound like you need direction. The interviewer is listening for what happens at the edges. When do you pull people in? How long do you stay stuck before you say so?
A preference is fine to state. It can’t be the whole answer. The useful part is the rule you follow when your preference stops working.
Companies that want independence still expect you to surface blockers
GitLab’s public handbook is a good example, because it states its expectation in writing. Its leadership page says it wants each team member to be “a manager of one.” Managers of one set their own priorities and decide what needs doing. That sounds like a vote for working alone.
The same page then lists what a manager of one does. One example is a team member who surfaces blockers instead of assuming the manager already knows. The handbook says they do it by “working in public and with a low level of shame.” Another example writes a proposal before asking for anyone’s time.
So even a company built on autonomy defines it with two handoffs. You bring a plan before you take people’s time. You say you’re blocked before anyone has to ask. Those are the two switch points your answer should name.
Name your default, then the point where you show the plan
Take Lena, an invented Android engineer with four years of experience. She works on a podcast app. She is interviewing for a senior role at a remote-first company.
Her honest default is solo work. She does her best work in long stretches with notifications off. Her last big task was moving episode downloads onto WorkManager. It took about five weeks, mostly alone.
She didn’t start it alone, though. Before writing code, she wrote a one-page plan. It covered how downloads would resume after the app was killed. It also covered what would happen to downloads already in progress. She posted it in the team channel and asked for objections within two days. A teammate pointed out that the car mode screen read download state directly. Her plan would have broken that screen. She changed the plan before it cost anything.
That is the first switch point. Her solo stretch begins after other people have seen the shape of the work. Say it in one sentence in the answer. Also say what the review caught, because that proves it wasn’t a formality.
Set a limit for being stuck, and admit when you learned it the hard way
The second switch point is harder to prove. “I ask for help when I need it” proves nothing on its own. The believable version has a limit you can name.
Intercom’s blog described one in a November 2021 post on the “15-minute rule.” The rule came from the company’s customer support engineering team. You spend 15 minutes trying to solve a problem any way you can. The post then states the rule’s second half. “If you don’t have an answer after 15 minutes, you must ask someone.” The post is from 2021 and about support tickets. A five-week project needs a longer limit. The idea transfers anyway. You decide in advance when stuck becomes a reason to talk.
Lena’s limit came from a mistake. Midway through the migration, some downloads finished with corrupt audio files. She spent three days alone on it. She checked her resume logic, then the file writes, then the retry policy. On the fourth day she finally posted the symptoms. A backend engineer recognized it in minutes. The CDN ignored range requests for one older host. So a resumed download restarted from zero while her code appended to the partial file.
After that, Lena set a rule. If a problem blocks her for half a working day, she posts what she has tried. That rule is the strongest part of her answer. It shows a real cost of her default and the fix she still uses. An interviewer can ask when she last used it. She has an answer ready.
Lena’s answer, start to finish
Lena, her app and the bug are invented for this post. Borrow the order and fill it with your own project.
My default is working independently. I do my best work in long, focused stretches. My last big project was moving our podcast downloads onto WorkManager. That was about five weeks, mostly on my own.
But I bring people in at two points. Before I start, I write a short plan and ask the team for objections. On that project, a teammate caught that car mode read download state directly. We changed the plan before any code broke.
The second point is when I’m stuck. I learned that one the hard way. Midway through, some downloads came back corrupt. I spent three days on it alone. When I finally posted the symptoms, a backend engineer knew the cause in minutes. One CDN host ignored range requests. Now if something blocks me for half a day, I post what I’ve tried.
So I’d say I work best alone between those two points. The team sees the plan first and hears about blockers early.
The answer opens with a real preference, so it doesn’t dodge. The middle paragraphs give the two switch points, each with evidence. The three lost days carry the most weight. They show Lena knows how her preference fails. The last line restates the rule in one sentence the interviewer can write down. For remote roles, the same habits answer the time zone question too.
Ask back how long the team waits to hear about a blocker
This question tells you as much about the team as about you. When Lena gets her turn to ask, she reuses her own rule. “When someone here is stuck, how soon would you expect to hear about it?”
Listen for a concrete reply. “Post in the channel within a day” or “raise it at standup” means the team has a norm. You can check your own limit against it. “We’re all pretty independent here” is a weaker reply. It may mean nobody notices when someone is stuck for a week. That’s the situation that cost Lena three days. Ask a follow-up about the last time a project slipped and when the team found out. Our post on asking why a role is open shows how to read vague answers like that one.