The questions you ask, by interviewer

What it is

The questions you ask are an evaluated part of every interview, and the same question asked of the wrong person wastes the slot and signals you do not understand the org. Asking a recruiter about the team's technical debt gets you a vague answer; asking the CTO about the interview timeline wastes 90 seconds with the person best placed to tell you where the company is going.

The principle: ask each person what only they know.

Recruiter       process, level, band, timeline, competing candidates
Hiring manager  the actual job, the team's problems, what success looks like
Peers           what it is REALLY like; the things the manager will not say
Future reports  (if leading) what they need; the previous manager's failures
Skip-level      strategy, the team's standing, resourcing, your manager
VP / CTO        direction, bets, what keeps them up at night
Product/design  how decisions get made; whether engineering is a partner
Bar raiser      the company's standards and how they are enforced

What this is confused with: questions as a formality at the end. They are an evaluation. A candidate who asks a hiring manager "what does success look like in six months" is demonstrating that they think in outcomes; one who asks "is there free lunch" has spent their signal.

And they are your due diligence. You are deciding too, and the interview loop is the only time you have this much access to the organisation. Wasting it means accepting an offer on brochure information.

The problem it solves

Two problems, and most candidates only recognise the first.

You are being evaluated on your questions. Interviewers report on them, and the difference between "do you have any questions?" answered with silence and answered with two sharp ones is a visible difference in the debrief.

You are making a decision with almost no information. The failure this prevents is the one people describe as "the job was not what I was told":

What the job description says:   "lead a team building our platform"
What you learn by asking:        the team is 3 people, 2 of whom are leaving,
                                 the previous lead left after 7 months, the
                                 platform is a rewrite that has slipped twice,
                                 and the VP who sponsored it has left.

Every one of those is discoverable by asking the right person the right question, and none of them appears in a job description. The reverse-due-diligence framing (see reverse due diligence) is the systematic version; this page is who to ask what.

Mechanics

The recruiter

They know process and constraints, and they do not know the technical details. Use them for things that are expensive to discover later.

"What level is this role mapped to, and what is the band?"
   -> ANCHOR EARLY. See level negotiation: the level determines the band,
      and it is far easier to establish before the loop than after.

"What does the loop look like, and who will I meet?"
   -> so you can prepare per-interviewer, which is this page's point.

"Is this a new headcount or a backfill?"
   -> a backfill invites "why did the last person leave?"

"What is the timeline, and are there other candidates at final stage?"
   -> real information about your leverage and their urgency.

Asking the level and band at the recruiter screen is the single highest-value question in the process, because the level is decided early and negotiating compensation inside a level you were mapped to low is a much weaker position (see level negotiation).

The hiring manager

The most important conversation, and the one where vague questions waste the most.

"What are the three biggest problems facing this team right now?"
   -> if they cannot name three, they have not thought about it,
      or they will not tell you.

"What does success look like at 3, 6 and 12 months?"
   -> a manager with no answer has no plan for you.

"Why is this role open?"
   -> growth, backfill, or a reorg. Each implies a different situation.

"What is the team's biggest constraint: people, clarity, or technology?"
   -> the answer tells you what the job actually is.

"How is the team's work prioritised, and who decides?"
   -> whether you will be executing someone else's list.

"What would make you regret hiring me?"
   -> uncomfortable, and it surfaces the real concern. The answer tells
      you what to address in your follow-up.

"Why is this role open" is the question with the highest information-per-word, and the follow-up matters: for a backfill, "what would have made that person stay?"

Peers

Peers tell you what the manager cannot. The manager is selling; the peer is your future colleague and has less incentive to.

"What is a recent week that went badly, and why?"
   -> concrete, and it defeats the rehearsed answer.

"How much of your time goes to unplanned work?"
   -> the honest measure of operational load.

"What is the on-call rotation, and when were you last paged at night?"
   -> the second half is the one that gets a real answer.

"How do decisions get made when two people disagree?"
   -> tells you whether the culture is consensus, authority, or attrition.

"What has changed since you joined, and was it for the better?"

"What would you change about how the team works if you could?"

"When were you last paged at night" is the question that produces an involuntary honest answer, because it is specific and recent, where "how is on-call?" produces "it's fine."

Future reports (for a lead role)

If you will manage them, this is the interview that matters most and the one candidates prepare least.

"What do you need from a lead that you are not getting?"
   -> the actual job description.

"What did the previous lead do well, and what would you have changed?"
   -> both halves. The second is the real answer.

"What is the most frustrating part of your week?"

"If you could stop doing one thing, what would it be?"

The reports are also evaluating you, and their feedback is often decisive for a lead role. Asking what they need signals that you understand the job is service rather than authority.

Skip-level

They know the team's standing in the organisation, which nobody else will tell you.

"How does this team's work connect to the company's priorities this year?"
   -> whether the team is central or peripheral. Peripheral teams get cut.

"What is this team's reputation with the rest of engineering?"

"How is headcount and budget decided, and what is this team's trajectory?"

"What is my future manager particularly good at?"
   -> a positive framing that produces a real answer, and the omissions
      are informative.

"What is this team's reputation" is the question that surfaces a team in trouble, and it is one a skip-level will answer honestly more often than you would expect, because they are recruiting you into a problem they want solved.

VP and CTO

The shortest slot and the highest-altitude information. Do not spend it on the team.

"What are the two or three bets the company is making this year?"

"What would have to be true in two years for you to consider this a success?"

"What is the biggest technical risk to the strategy?"

"What is the thing you believe about this business that most people
 outside would disagree with?"

Ask about direction, not about your team, because they know direction and someone else knows your team. The CTO page (see 90 seconds with the CTO) covers the compressed version.

Product and design partners

They tell you how decisions actually get made, which is the thing most likely to determine whether the job is enjoyable.

"How does a feature go from idea to shipped here? Walk me through a recent one."
   -> a real example defeats the idealised process description.

"When engineering says something will take three months and product wants
 six weeks, what happens?"
   -> the honest answer to "are we partners or a service function."

"What do you wish engineering understood better about your work?"

"How do you decide what NOT to build?"

"Walk me through a recent one" is the technique: a concrete recent example is much harder to sanitise than a description of the process.

The bar raiser

They are enforcing a standard, so ask about the standard.

"What separates someone who is hired at this level from someone who is not?"
"What is the most common reason strong candidates are not hired here?"
"How does the company keep the bar consistent as it grows?"

The three universal closers

Ask these of nearly everyone; the variation in answers is itself the signal.

1. "What is the biggest challenge facing the team/company right now?"
   -> if five people give five different answers, there is no shared
      understanding of priorities. That is data.

2. "What do you like most about working here, and what would you change?"
   -> the second half. A long pause is informative; so is a rehearsed answer.

3. "Is there anything about my background that gives you hesitation?"
   -> THE MOST VALUABLE QUESTION IN THE LOOP. It surfaces the objection
      while you can still address it, rather than losing to it in the debrief.

Closer 3 is the one candidates avoid and the one that changes outcomes. An unstated concern becomes a "no" in the debrief that you never hear about; a stated one is something you answer in the room or in the follow-up.

Debrief dynamics, and the follow-up

Understanding what happens after helps you aim the follow-up.

The debrief:
  - each interviewer writes independently BEFORE the meeting (at good
    companies), to avoid anchoring
  - the loop meets; the hiring manager and bar raiser carry the most weight
  - "no hire" is easier to defend than "hire": one confident detractor can
    sink a candidate that four people liked
  - the DECIDING factor is often a specific unresolved concern, not an
    overall impression

The follow-up that lands is the one that addresses the concern.

Generic:  "Thank you for your time, I enjoyed our conversation."
          -> read, ignored, changes nothing.

Effective: "You asked about how I would handle X. I have been thinking about
           it, and here is the approach I would take: [3 sentences]. It is
           the same shape as [a thing you actually did]."
          -> gives your advocate ammunition in the debrief.

Send it within 24 hours, because the debrief happens fast. And send it to the person whose concern it addresses, or to the recruiter with a request to pass it on: an email arriving after the decision is worthless.

If you asked closer 3, you know what to write. That is why it is worth the discomfort.

A worked example: a loop where the questions changed the outcome twice

A staff engineer interviewing at a mid-size company for a platform lead role.

Recruiter screen:

Q: "What level is this mapped to, and what is the band?"
A: "It's an L5, band is $X-$Y."

The candidate's research suggested the scope described was L6 at this
company. They raised it THEN:
  "The scope you described (owning the platform roadmap, three teams
   consuming it) sounds like what your L6 job architecture describes.
   Can we confirm the level before I invest in the loop?"

Result: the recruiter checked, and the role was re-posted at L6.

Raising the level at the recruiter screen, before any technical interview, changed the band by about 35 percent, and it was possible because the level had not yet been anchored by the loop. After the loop the same conversation is "we assessed you at L5," which is far harder to move (see level negotiation).

Hiring manager:

Q: "Why is this role open?"
A: "The previous lead moved to another team internally."

Q: "What would have made them stay in this role?"
A: (pause) "Honestly, I think they found the stakeholder management
    exhausting. There are four teams with different priorities and
    no clear arbitration."

The follow-up produced the real answer, and it reframed the job: the technical work was not the hard part, the arbitration was. The candidate spent the rest of the loop testing that hypothesis rather than the platform's architecture.

Peer interview:

Q: "When were you last paged at night?"
A: "Tuesday. And the Tuesday before."

Q: "What was it?"
A: "The same thing both times. We know what it is, it's on the backlog,
    it's been there six months."

A known, recurring, unfixed page is a resourcing and prioritisation signal, not a technical one, and it corroborated the hiring manager's arbitration answer: the team could not get priority for its own reliability work.

Skip-level:

Q: "What is this team's reputation with the rest of engineering?"
A: "Mixed. They're seen as slow, which I think is unfair given what
    they're being asked to absorb."

Q: "How does the team's work connect to this year's priorities?"
A: "It's an enabler for two of the four company goals."

"An enabler for two of four goals" plus "seen as slow" is a specific, addressable situation, and it is exactly what a platform lead is hired to change. It also means the team is not peripheral, which is the risk with platform roles.

The closer, with the hiring manager:

Q: "Is there anything about my background that gives you hesitation?"
A: "You've worked at larger companies. I'd want to know you can operate
    without the support structures you're used to."

A specific, addressable concern, stated out loud. The follow-up email, sent that evening:

"You mentioned the concern about operating without larger-company support
 structures. Two things: at [company] I built the on-call rotation and the
 runbook set for a team that had neither, and before that I was the second
 engineer at [smaller company] where there was no platform team at all. If
 it is useful, I can walk through how I would approach the first 90 days
 here given the arbitration problem you described, which seems like the
 harder half of the job."

The email did two things: it answered the stated objection with two specific counter-examples, and it demonstrated that the candidate had correctly identified the real problem (arbitration, not technology) from the earlier answer.

The outcome:

offer:  L6, at the level established in the recruiter screen
the hiring manager later said the follow-up email was decisive in the
debrief, because it was the only evidence against the one stated concern.

The questions changed the outcome twice: the level question changed the band before the loop began, and the hesitation question surfaced the objection in time to answer it.

And the candidate's own decision was made on information no job description contains: a team with a real arbitration problem, a known unfixed reliability issue, a "slow" reputation the skip-level thought was unfair, and a role that enables two of four company goals. That is a job you can decide about.

Production evidence

Amazon's Bar Raiser program is the best-documented example of a dedicated interviewer enforcing a hiring standard independent of the team's urgency, and the role exists precisely because a hiring manager under pressure will lower the bar.

Written-before-meeting debrief practice (independent written feedback before the group discusses, to avoid anchoring) is documented by Amazon, Google and others, and it is why your follow-up must arrive before the debrief rather than before the decision.

The asymmetry that "no hire" is easier to defend than "hire" is a consistent feature of structured hiring processes, and it is why an unresolved concern from one interviewer is disproportionately damaging. That asymmetry is the argument for closer 3.

Level-first negotiation is standard advice from compensation negotiation practitioners (Levels.fyi's guidance, Haseeb Qureshi's writing), and the mechanism is that the level determines the band, so negotiating within a band you were mapped into low has a much lower ceiling.

Reverse due diligence as a named practice appears in engineering-leadership writing (Charity Majors and others), with the argument that the interview loop is the highest-access window you will ever have into an organisation you are considering joining.

The debate

How many questions should you prepare? More than you will ask, targeted per interviewer. Three or four per person, prepared, plus the universal closers is the shape that works, because slots run short and you want the highest-value question first. Preparing one list and asking it of everyone is the failure, because it wastes the slots where only that person knows the answer.

Should you ask uncomfortable questions? Yes, and closer 3 in particular. The asymmetry is that an unstated concern becomes a "no" you never hear about, and a stated one is something you can answer. The discomfort is a few seconds and the information is decisive.

Is asking about compensation at the recruiter screen too early? Asking about the level and band is not too early, it is the right time, because the level is anchored early and moved with difficulty later. Asking for a specific number too early is a different question and is worth deferring; asking what level the role is mapped to is process information the recruiter holds.

Do peers give honest answers? More honest than the hiring manager, and the technique matters: specific and recent beats general. "When were you last paged at night" produces a real answer; "how is on-call" produces "it's fine." The same applies to "walk me through a recent feature" versus "how does your process work."

Does the follow-up email matter? Only if it addresses something specific. A generic thank-you changes nothing. An email that answers a stated concern gives your advocate ammunition in a debrief where one confident detractor can sink you, and in the worked example the hiring manager said it was decisive. It must arrive within 24 hours, because the debrief is fast.

What if you are asked "do you have any questions?" and you have already had them answered? Say so, specifically: "you covered the team's priorities and how decisions are made, so the one left is [X]." That demonstrates you were listening and had prepared, which is better than manufacturing a question and much better than "no, I think we covered everything."

Follow-up Q&A

"How do you decide what to ask whom?"

Ask each person what only they know. A recruiter knows process, level and timeline; a hiring manager knows the team's actual problems and what success looks like; peers know what it is really like day to day; a skip-level knows the team's standing and resourcing; a VP or CTO knows direction and bets; product partners know how decisions actually get made. Asking a recruiter about technical debt or a CTO about the interview timeline wastes the slot and signals you do not understand the org.

"What is the single highest-value question in the loop?"

Two, at different points. At the recruiter screen: "what level is this mapped to, and what is the band," because the level is anchored early and determines the band, and moving it after the loop means arguing against an assessment. At the end of each interview: "is there anything about my background that gives you hesitation," because an unstated concern becomes a "no" you never hear about, and a stated one is something you can answer in the room or in the follow-up.

"How do you get honest answers from peers?"

Specific and recent, rather than general. "When were you last paged at night" gets a real answer where "how is on-call" gets "it's fine." "Walk me through a recent feature from idea to shipped" defeats the idealised process description. "What is a recent week that went badly" is harder to sanitise than "what are the challenges." The pattern is that a concrete recent instance is much harder to answer with a rehearsed line.

"What makes a follow-up email effective?"

Addressing a specific stated concern, within 24 hours. A generic thank-you changes nothing. If you asked whether anything gives them hesitation, you know exactly what to write, and the email becomes evidence your advocate can use in a debrief where one confident detractor can sink a candidate four people liked. In one case the hiring manager said it was the only evidence against the single stated concern and it was decisive.

"Why ask about level before compensation?"

Because the level determines the band, so negotiating compensation inside a level you were mapped to low has a low ceiling. And the level is anchored early: raising it at the recruiter screen is a process question, while raising it after the loop means arguing against an assessment several people contributed to. In one case raising it at the screen, with a specific comparison to the company's own job architecture, moved the role from L5 to L6 and the band by about 35 percent.

"What do the questions tell you that a job description does not?"

Everything that determines whether the job is what you were told. Why the role is open and what would have made the previous person stay; how much time goes to unplanned work; whether a recurring page has been unfixed for six months; the team's reputation and whether its work connects to company priorities. None of that appears in a description, and the loop is the only time you have that much access.

Common misconceptions

"Questions are a formality at the end." They are evaluated and reported on, and they are your only due diligence window. A candidate who asks a hiring manager what success looks like at six months is demonstrating outcome thinking.

"Ask the same questions of everyone." It wastes the slots where only that person knows the answer. The universal closers are the exception, and the variation in their answers is itself signal.

"Do not ask about compensation early." Asking the level and band at the recruiter screen is correct timing, because the level anchors early and determines the band. A specific number is a different question.

"A thank-you email is polite and harmless." A generic one changes nothing. An email answering a stated concern is evidence in a debrief where a single unresolved objection is disproportionately damaging.

"Asking if they have hesitations looks weak." It surfaces the objection while you can still answer it. The alternative is losing to a concern you were never told about.

Interview delivery note

Say this verbatim: "Ask each person what only they know, and end with 'is there anything about my background that gives you hesitation.' That last one is the highest-value question in the loop, because an unstated concern becomes a no you never hear about, and a stated one is something you can answer in the room or in the follow-up email." The organising principle and the single question that changes outcomes.

The senior-versus-staff separator is raising the level at the recruiter screen. A senior candidate prepares good questions for the technical interviews. A staff candidate knows the level is anchored before any technical conversation, that it determines the band, and that the recruiter screen is the moment to compare the described scope against the company's own job architecture. Moving from L5 to L6 there is a different negotiation from moving it afterwards, when you are arguing against an assessment several people contributed to.

The second signal is asking peers specific, recent questions. "When were you last paged at night" producing "Tuesday, and the Tuesday before, same thing, it has been on the backlog six months" is information no general question extracts, and it is the difference between due diligence and polite conversation.

Further reading

• Amazon's published description of the Bar Raiser role, for how a hiring standard is enforced independently of a team's urgency.
• Haseeb Qureshi, "Ten Rules for Negotiating a Job Offer," particularly on establishing level before compensation.
• Charity Majors' writing on interviewing the company, for the reverse-due-diligence framing.
• The reverse due diligence and level negotiation pages in this chapter, which this page's question set feeds.