The best Principal Engineers I've worked with were not the smartest people in the room. They were the people who made the room smarter.
This is the lens I use when I interview. It is also why I no longer ask system-design improv questions.
You are not hired to be the smartest person in the room. You are hired to make the room smarter.
Thermostats, not heroes
A heroic engineer fixes the outage at 3am. A Principal designs the system so the outage doesn't happen — and if it does, three other engineers can fix it without paging anyone. The job is to set the temperature of the room, not to be the warmest body in it.
Heroism scales linearly with the hero. Thermostats scale with the team. One of those is a sustainable career; the other is a path to burnout that the org will applaud right up until you leave.
What I actually ask
Three questions, in roughly this order:
- Tell me about a decision your team made that you disagreed with — and lost. What did you do next?
- Walk me through an ADR you wrote that changed your team's behaviour six months later.
- Where does your influence stop, and how do you know?
Each one is designed to surface a specific quality. Question one separates engineers who can disagree-and-commit from those who carry resentment into execution. Question two separates writers from talkers. Question three separates self-aware seniors from people who think their reach is unlimited.
What I avoid
I no longer ask whiteboard system-design questions. They reward fluency in performance theatre — the candidate who has memorised "throw a queue at it" and "shard by user_id" wins, regardless of whether they can navigate the actual mess of a real system.
Instead I bring real artefacts to the interview. A messy ADR I wrote three years ago, with the wrong call in section 4. "Tell me what you'd push back on." The artefact does the screening for me. People who can read it well — who find the soft spot in a real document — are the people I want to work with.
Signals I watch for
- Names other engineers without prompting. The ones who say "my colleague Jane figured this out" are the ones who'll grow other people on your team.
- Disagrees with their past self in the interview. Self-revision is a strong indicator of intellectual honesty.
- Asks about constraints, not just goals. "What's the hardest part of your platform right now?" beats "What does success look like?" every time.
Anti-signals
Talks only in first person plural about successes ("we shipped") and first person singular about failures ("I was in a tough spot"). Treats every story as a victory. Cannot give a straight answer to "what's a system you wish you hadn't built." These aren't dealbreakers, but they're cumulative — three or four of them and I know what kind of teammate I'm hiring.
Closing thought
Hire the thermostat. The hero will arrive eventually anyway, and the room will already be the right temperature when they do.
— Junaid · 30 July 2025