For years the questions round was the part of the interview nobody owned. The recruiter told you “save ten minutes at the end for your questions,” the interviewer half-listened, and whatever you asked never made it into the debrief (the meeting afterward where the panel compares notes and decides). That’s changing at the more deliberate teams. The block gets named, it gets real time, and what you ask shows up in the write-up.
At the teams that run it on purpose, it looks the way I’ve seen it run: a dedicated 30 to 45 minutes late in the final-round day, often with the hiring manager or a peer you haven’t met yet, placed there so you’ve seen enough of the team to ask something informed. The other change is mechanical. AI note-takers now transcribe the whole conversation, questions included, and drop it into the evidence record the panel reviews. So the questions that used to evaporate the second you left the room increasingly get captured in the debrief and read as signal.
Be clear about how common this is. A questions round that’s formally scored against a rubric is still the exception, run by teams that are deliberate about hiring, not the default everywhere. You can usually spot it: the round is named ahead of time and given a real block of 30 to 45 minutes rather than the leftover ten. Prepare the same way regardless. Even on an ordinary “ten minutes at the end” loop, a sharp interviewer forms an impression from what you ask, so the by-round advice below is worth following whether or not anyone writes a number down.
Why a team would bother scoring your questions
Because the questions are cheap signal on the things that are hardest to test directly. A take-home tells you whether someone can write code. It tells you almost nothing about whether they’ll ask the right question before building the wrong thing for three weeks. In the debriefs I’ve sat in, the questions round comes up as a tiebreaker far more often than anyone admits in the rubric: a senior candidate who spent the whole slot on PTO and remote policy reads differently than one who wanted to know what the last person in the role found hardest.
At shops that run a weighted scorecard, you’ll often see four to six signals, each on a scale with a line of quoted evidence, which is exactly the shape that makes “quality of questions” easy to bolt on. Others keep it simpler: a 4-point hire / no-hire read with a comment box. Either way the evidence is right there in the transcript. You don’t need a new round to score it. You need one interviewer to write “asked how on-call is staffed and what broke most recently, clearly probing load not comp” into the notes, and now it’s a data point.
What strong questions to ask the interviewer sound like, by round
The single most common mistake is asking everyone the same three questions. Each person in the loop (the full set of interviews you go through) can answer something no one else can, and asking the recruiter a deep architecture question (or the staff engineer about the PTO policy) signals you don’t know who you’re talking to. Match the question to the role.
Recruiter: process, money, and reality checks
The recruiter owns the pipeline and the offer, not the engineering. Use them for the things they actually control and the things they’ll answer straight because it’s their job to keep you warm. On the recruiter’s side, the recruiter phone screen is scoring your motivation and your read on the role, so your questions land best when they show you’ve thought about the job rather than just the paycheck. Good ones:
- “How many people are in the loop after this, and what’s each one weighting?”
- “What’s the band for this level (the pay range the role is slotted into), and is the number flexible on base, equity, or both?” — recruiters expect this, and it’s the right room for it; the heavier parts belong in a real salary negotiation once you have the offer.
- “Is this a backfill (replacing someone who left) or new headcount (an added role), and why is it open?”
- If it applies to you: “Does this role sponsor a visa, and is it open to my location, or is the band adjusted by where I’m based?” Worth settling early, before four more rounds go into a dealbreaker.
Weak with a recruiter: “What does your tech stack look like?” They’ll recite the job post back to you, and you’ve burned a question on something you could have read.
Hiring manager: the actual job and how they run the team
This is the person you’ll report to, so the scoring here is the heaviest. They’re listening for whether you understand what the role is for. Aim at the work and the tradeoffs:
- “What’s the thing that’s broken right now that you’re hoping this hire fixes?”
- “When your team and another team disagree on a design, how does that actually get resolved?”
- “What did the last person who did well in this role do in their first six months?”
If you’re early-career or interviewing for a new-grad cohort, that last one doesn’t map, there was no previous person in your seat. Swap it for something scaled to where you are: “What will I be trusted to own by the end of my first year?” or “What does a new grad who’s ramping well look like here after six months?”
Weak: “What’s your management style?” It invites a rehearsed answer and tells the manager nothing about you.
Peer engineers: what the day really feels like
Peers will tell you things the manager won’t, and a sharp peer question signals you know where the unfiltered information lives. They’re also quietly assessing whether they want you in their standup.
- “What’s the worst part of the codebase, and is anyone allowed to fix it?”
- “Walk me through how a change goes from your laptop to production. How long does that take?”
- “What’s something you wish you’d known before you joined?”
Weak: anything that’s really a brag in disguise (“How do you handle engineers who are just too fast for the team?”). Peers clock it instantly.
In an engineering-manager loop this round shifts, because your “peers” are other managers, not the engineers you’d lead. The signal they’re reading is cross-team partnership, not whether they’d want you in their standup. Ask them how managers divide ownership when projects span teams, or where the friction between engineering and product currently sits.
Skip-level or exec: direction and whether the bet is sound
With a skip-level (your manager’s manager) or a founder, scope up. Asking them about your onboarding buddy wastes the one person who can tell you where the company is going. Match the altitude:
- “What has to be true in eighteen months for this to have been the right bet?”
- “Where do you and the leadership team currently disagree?”
- “What’s the thing about this business that’s harder than outsiders assume?”
Early-career readers: scoping up doesn’t mean grilling a founder on board politics you can’t evaluate. “What do you want someone in my role to be great at two years from now?” or “What’s the bet this team is making that you’re most confident about?” reaches the right altitude without pretending to a seniority you don’t have yet.
Weak: a logistics question. It reads as not understanding why you’re in the room with them.
| Round | Who you’re talking to | What they’re reading in your questions | A strong question | A weak question |
|---|---|---|---|---|
| Recruiter screen | Pipeline and offer owner | Whether you’re serious and know how hiring works | “Is this a backfill or new headcount, and why is it open?” | “What’s your tech stack?” |
| Hiring manager | Your future boss | Whether you understand what the role is for | “What’s broken now that you need this hire to fix?” | “What’s your management style?” |
| Peer engineer | Future teammate | Whether you dig for real signal and would be good to work with | “How long does a change take to reach production?” | “How do you handle engineers who are too fast for the team?” |
| Skip-level / exec | Direction setter | Whether you think about the whole business, not the task alone | “What has to be true in 18 months for this to be the right bet?” | “How does onboarding work?” |
The questions that cost you points
Googleable questions are the fastest way to lose the round. If the answer is on the careers page, the funding announcement, or the first paragraph of the job description, asking it tells the panel you didn’t prepare. The fix isn’t to avoid the topic, it’s to show you did the reading and want the part that isn’t public: not “did you raise a Series B,” but “you raised in the spring, so how has headcount planning changed since then.”
Pure perks questions early read badly too, not because wanting good PTO is wrong, but because sequencing it first signals it’s your top filter. Save those for the recruiter and the offer stage. Avoid the compound question that’s really five questions stapled together, because the interviewer answers the easy one and you lose the rest. And skip the questions you don’t actually want answered, the ones you memorized off a list, because the follow-up exposes them fast. If you ask what the biggest technical challenge is and then have nothing to say when they tell you, the question worked against you.
How a team that scores this decides
When a team grades the questions round, the rubric tends to reward four things, and you can aim at all of them without gaming anything. First, listening: a question built on something an earlier interviewer said shows you were paying attention across the loop instead of reading from a card. Second, priorities: what you probe first is taken as what you’ll optimize for on the job, so leading with the work rather than the perks matters. Third, judgment: the best questions reveal how you’d make a decision, which is the same muscle a good behavioral answer is trying to show, just pointed outward. Fourth, scope that matches your level: a staff candidate asking only about their own ramp reads as under-leveled, while a new grad grilling the CEO on board dynamics reads as the opposite.
This is a different axis than the AI-screening and automated-round changes reshaping the early funnel, which the AI-era interview guide covers on its own. The reverse interview is still a human reading a human. The only thing automation changed is that the transcript now remembers what you asked.
The useful reframe: you were going to use this time to decide whether you want the job anyway. The questions that actually help you decide, the ones about how the team really works and what you’d be walking into, are the same ones that score well. You don’t have to choose between asking for yourself and asking for the panel. For more on preparing each stage, the interview prep guides go round by round. Ask like someone who might say no.
Practice the behavioral round:
