The Interviewer Cue Reference, Each In-Round Move Keyed to Its Rubric Line
Look up any of 22 interviewer moves mid-round and know the rubric axis it feeds, the misreading to avoid, the reply that earns an evidence line, and what to log.
Key takeaways
- In three public coding rubric reconstructions, solving with hints sits in the Lean Hire band and needing significant hints sits in Lean No Hire, so one hint you use quickly costs less than repeated hints or a hint you ignore.
- Published advice on when to ask for a coding hint ranges from about 3 minutes stuck to 8 to 10 minutes of real effort, roughly a 3x spread, so set your own threshold before the round instead of guessing in it.
- OPM's structured interview guidance fixes the range of probes before the interview and asks interviewers to use very similar probes for every candidate, so a situation, action, or result probe usually points to a gap in your answer, not suspicion of you.
- "Let's move on" and "I got the signal" are read both ways in interviewer accounts, so the useful check is whether you had already stated a working approach before the cut, not the phrase itself.
- Interviewers at mid-to-large tech companies typically write independent evaluations before a calibration debrief, which means the written hint log, not the mood of the room, is what gets compared.
- In Refolk's index, 17 US and 10 Indian profiles hold Technical Interviewer or Interview Engineer titles, with Karat the top employer in both, so standardized probe and hint ladders are a plausible setup to prepare for.
This is a lookup table for the moment inside a live interview round when the interviewer hints, redirects, challenges, probes, warns about time, goes quiet, or wraps early. It is for candidates in coding, case, behavioral, and panel rounds. For each of 22 moves it gives the scorecard axis the move most often feeds, the common false reading, the reply that produces a recordable evidence line, and the one fact to log afterward.
Two terms recur throughout. A scorecard axis is a named dimension the interviewer rates, such as Problem Solving or Communication. An evidence line is a concrete observation the interviewer writes in their notes and later uses to justify a rating, for example "received a hint on using a hash map but did not apply it." This guide covers moves inside the round only.
What an interviewer move is actually grading
Most in-round moves are planned tools for collecting evidence against a fixed set of rubric axes. They are not running commentary on how the interviewer feels about you. Read the move as a question about one axis, and your reply as the evidence line that axis gets.
The clearest public description of this is OPM's Structured Interviews: A Practical Guide, dated September 2008. It tells interviewers to set the range of probing before the interview: no probes, a limited number, or unlimited. Where probes are used, it tells them to decide in advance which specific probes are allowed for each question. Interviewers should "use very similar probes for all candidates to ensure candidates are given the same opportunities," and OPM's training slides add "Do not use leading probes that convey the answers." A structured interview typically assesses 4 to 6 competencies.
Google's re:Work guide describes the same shape: questions "often contain an initial prompt, along with several follow-ups," and predetermined follow-ups "help elicit a high level of detail from the candidate."
The practical consequence: when a follow-up arrives, the most likely meaning is that your answer has not yet supplied the detail a rubric line needs.
The six move types
Before you answer anything, put the move into one of these buckets:
- Hint: the interviewer gives you information you did not have.
- Redirect: the interviewer moves you off what you are doing.
- Challenge: the interviewer questions whether you are right.
- Probe: the interviewer asks for more detail on what you already said.
- Time cue: the interviewer tells you about the clock, or goes silent.
- Wrap: the interviewer closes the round or invites your questions.
One vendor further sorts coding hints into four kinds: complexity probe, edge case nudge, data structure steer, and redirect via example. A hint ladder is the sequence of increasingly specific hints an interviewer can give. Those four kinds are its usual rungs.
Coding hints: the six hint moves and what each grades
Hints in coding rounds feed the Problem Solving axis in every public reconstruction. What hurts you is needing significant help or failing to use the hint, not receiving one. The table covers each hint move. The log line for each follows it.
| Move | False reading | Reply that earns a line |
|---|---|---|
| Complexity probe ("what's the runtime?") | My approach is wrong | State the complexity, then say if and how you would improve it |
| Edge case nudge ("what if it's empty?") | I missed something fatal | Trace the case aloud, fix the code visibly |
| Data structure steer ("could a map help?") | The hint sank the round | Restate what the structure buys, change the code within a minute |
| Redirect via example ("try this input") | I'm being rescued | Run the example aloud, name what it reveals |
| Hint you asked for | Asking looks weak | Ask narrowly: what you know, where blocked, proposed direction |
| Hint offered while you were still thinking | I already failed | "I was heading toward X, let me check it against that" |
What to log for each hint move:
- Complexity probe: the complexity you stated and whether you changed the approach.
- Edge case nudge: which case, and whether your code handled it after the fix.
- Data structure steer: the minutes from hint to working code.
- Redirect via example: what the example exposed in your approach.
- Requested hint: how long you had been stuck and the exact wording you used.
- Offered hint: roughly how many seconds you had been thinking before it came.
The last row matters more than it looks. One Google L5 candidate reported a round rated Lean Hire after the interviewer hinted "while I was thinking just for 5 seconds." Another wrote: "I needed a hint for a bug - fair enough he did give me a hint, but I didn't ask for it." Interviewers note "whether you asked for help, what hint was given, and how fast you used it." Whether an offered hint and a requested hint are recorded differently is not established publicly. So log which one it was, and note that you cannot control the first kind.
How hint use maps to rating bands
Three public sources describe the same gradient. Each column below is taken from one source.
| Band | copilotinterview | GitHub rubric | interviewcrunch |
|---|---|---|---|
| Hire | minor gaps | working solution, communicated clearly | did well |
| Lean hire | got there with hints | solution worked with hints | solved but needed a hint |
| Lean no hire | needed significant help | needed significant hints | hesitant |
| No hire | could not make progress | no meaningful progress | not ready / significant issues |
All three separate "with hints" from "significant" help, and that boundary is the one to manage. A self-described interviewer put it plainly: "Hints are not deal breaker at all. Spoon feeding is." Two qualifiers apply:
- Seniority moves the bar. A former hiring manager says heavy hints alone can drop a senior candidate from Hire to Lean No Hire, because finding the path independently is what senior rounds test.
- Bands add up across the loop. One guide puts it as "One lean no hire can be outweighed by strong rounds elsewhere, but two or more usually can't."
When to ask for a hint
Published thresholds disagree by about 3x, so decide yours before the round.
| Source | Threshold |
|---|---|
| interviewskool | 3 min stuck; 5 min on one bug |
| codeintuition | 3 to 5 min visible effort |
| nostrobeai | "a couple of stuck minutes" |
| techscreen | 8 to 10 min real effort |
| IGotAnOffer | last resort; max 1 to 2 hints |
The sources also disagree on order. IGotAnOffer treats asking outright as a last resort that comes after you narrate your confusion. Pair-programming guides treat asking directly as normal collaboration. They agree on the form: show your thinking first, then ask a narrow question. My default is the middle of the range, roughly 5 minutes of visible effort. In a round explicitly framed as pair programming, I would ask sooner.
Here is where I am: [what I know works, e.g. the brute force is O(n^2) and correct]. I am blocked on [the one specific sub-problem, e.g. avoiding the inner loop]. My best guess is [proposed direction]. Am I on the right track, or should I look elsewhere?
Fill the brackets with your actual state; keep it to three sentences so the interviewer can answer with one rung of the ladder.
Redirects, cutoffs, and "I got the signal"
A redirect means the interviewer wants your time spent somewhere else. It is a bad sign only when it arrives before you had stated a working approach, or when you resist it. Interviewer accounts genuinely conflict on what "let's move on" means, so read your own state, not the phrase.
| Move | Usually feeds | False reading | Reply that earns a line |
|---|---|---|---|
| "Let's move on" mid-solution | The question's own axis | Always bad | One sentence: approach, what's left; then move |
| "I got the signal I was looking for" | Signal already scored | Always good, or always bad | Same one-line summary; don't relitigate |
| Cut off mid-answer | Communication | They're bored with me | Give the headline in one sentence, offer detail |
| Redirect off a wrong sub-problem | Problem Solving, decomposition | They think I'm lost | Drop it, name the new sub-problem, act |
What to log for each redirect:
- "Let's move on": whether you had already stated a working approach.
- "I got the signal": the minute it came and what was on screen.
- Cut off: how long you had been talking and whether the question had a narrower ask.
- Sub-problem redirect: whether you recovered and how long it took.
On "I got the signal," one interviewer uses it "if I see that a candidate is stuck," and a second account says the same. So it is not a reliable compliment. On cutoffs, a hiring manager says interrupting long answers happens when "the candidate is not giving succinct answers," which is why I map it to Communication.
How you take a redirect is itself recorded. One vendor says "candidate argued when redirected" feeds "directly into both Problem-Solving and Communication scores." Another files "Got stuck on the wrong sub-problem and could not recover even when redirected" under problem decomposition. "Hard to redirect" as an exact rubric phrase is not established publicly, but the behavior clearly is scored.
Challenges, "what else?", and extension questions
A challenge tests whether you can weigh evidence under pressure. The strong signal is that you "disagree with evidence," not that you hold or fold. "What else?" tests whether your structure survives being extended, and an extension question tests depth on something you have already built.
| Move | Usually feeds | False reading | Reply that earns a line |
|---|---|---|---|
| "Are you sure?" | Problem Solving; Composure in finance cases | I'm wrong, so change it | Trace or test, then defend or update |
| Planted wrong hint | Judgment on teammate input | Accept every hint | Evaluate it aloud; accept or decline with a reason |
| "What else?" (case) | Structure | More ideas score more | Add ideas inside named buckets |
| Extension question ("now what if...") | Depth | A trap, or a reward | Restate the new constraint, say what breaks |
What to log for each challenge:
- "Are you sure?": whether you held or updated, and what evidence you used.
- Planted hint: whether you accepted it, and your stated reason.
- "What else?": how many buckets you named and whether new ideas stayed inside them.
- Extension: the new constraint and whether your solution needed rework.
The matrix below covers the "are you sure?" decision. The axes are whether you produced evidence and whether you changed your answer.
Answering "are you sure?"
Case rubrics make the "what else?" point directly. First-round interviewers "rate your presence, structure, depth, use of a hypothesis, and how well you drive the case." On extension prompts, "A list of random ideas scores poorly, even if some ideas are good." The Management Consulted scoring rubric checks whether ideas stay categorized when "what else" is asked. A UK finance case guide lists "defensive when challenged" as the weak signal under Composure.
Weighting is less certain. One prep site says fit is "roughly 50% of the final round scoring" in consulting, but firm-official weights are not established publicly.
For extension questions in coding rounds, no published rubric names a specific axis. Treat the question as one of the "several follow-ups" re:Work describes. It is planned, not a verdict.
Behavioral probes on situation, action, and result
A behavioral probe is a pre-planned follow-up that fills in a missing piece of your story: the Situation or Task, the Action you took, or the Result. Answer it directly first, in one or two sentences, and do not retell the story from the top.
OPM groups probes by Situation/Task, Action, and Result. Its situational table uses "Outcome" for the last column. A secondary analysis says OPM's one-page example pairs one lead question with three probes, while its probe table offers nine or ten per question. That count is secondhand, so check the PDF if it matters to you. One vendor says Google behavioral rounds run 45 minutes with "two to three main behavioral questions with significant follow-up probing on each."
| Move | Usually feeds | False reading | Reply that earns a line |
|---|---|---|---|
| Situation/Task probe ("what was your role?") | The lead question's competency | They doubt the story | One sentence of scope: team, stakes, your remit |
| Action probe ("what did you do?") | Same competency | I was too vague overall | "I" statements: the decision, the step, the tool |
| Result probe ("what happened?") | Same competency | The story failed | The outcome, measured if possible, then what changed |
| Situational Outcome probe ("what if it wasn't received well?") | Same competency | A trick question | One concrete next step and why |
What to log for each probe:
- Situation/Task: which story and which detail was missing.
- Action: whether you said "we" where the probe wanted "I."
- Result: whether you had a number or concrete change ready.
- Outcome probe: the scenario they posed and the step you named.
Because OPM fixes probes in advance and asks for the same ones across candidates, a probe usually shows a gap in your answer rather than suspicion of you. If your log shows the Result probe landing on the same story round after round, the fix is upstream, in how that story is written down. Refolk writes your resume from your own history and tailors it to each posting, so the lines an interviewer reads before the round already carry the result a probe will ask for.
Time warnings, silence, early wraps, and closing questions
Time cues and wraps are the moves candidates over-read most, and they have the least public rubric evidence behind them. By itself, an early end is neutral. Prioritize out loud on a time warning, narrate through silence, and use the close for prepared questions.
| Move | Usually feeds | False reading | Reply that earns a line |
|---|---|---|---|
| Time warning ("five minutes left") | No public rubric line | I'm out of time, so I've failed | Say what you'll finish and what you'll skip |
| Silence | No public rubric line | They've given up on me | Narrate your current step; ask if they want you to continue |
| Early wrap | Not established | Rejection | Use the time for prepared questions |
| "Any questions for me?" | Not established | A chance to ask how I did | Prepared questions; no "how did I do?" |
What to log for each:
- Time warning: the minute it came and the state of your solution.
- Silence: roughly how long, and what you were doing during it.
- Early wrap: minutes early, and whether you were invited to ask questions.
- Closing: anything said about next steps, word for word.
I found no public rubric that grades silence or time warnings directly. The replies above follow from what every hint source agrees on: show your thinking, because the interviewer can only write down what they can see.
On early endings the evidence is thin and contradictory. One AI-prep vendor asserts that ending "about 30% earlier than scheduled" is a strong negative sign but gives no methodology. A poster on an interviewer forum reads early endings as "more likely a sign of doing well or doing exceptionally badly." A careers site calls it neutral on its own. None of these is outcome data.
Do not ask how you did, because "final decisions are usually made by a panel afterward." The flow below shows where your round goes next.
Where an in-round move ends up
- The roomThe interviewer hints, probes, or redirects, and you reply
- NotesThe reply becomes an evidence line, such as how fast you used a hint
- Independent write-upEach interviewer submits an evaluation before seeing anyone else's
- Calibration debriefInterviewers compare written evaluations as a group
- DecisionTypically within 24 to 72 hours of the final interview
That timing and format come from a practitioner who has sat in hundreds of debrief calls, describing mid-to-large tech companies. How often a debrief overrides how a round felt is not established publicly.
How in-round readings go wrong
The most common mistake is reading a single move as a verdict on the whole round. Every failure mode below has a check you can run on your own log, and the check is always about what you did next, not what the interviewer said.
- Hint read as failure. You assume an offered hint sank the round. Check whether you turned it into code within a minute. The damaging note is "could not use the hint," not "got a hint." Vendor examples like "received a hint on using a hash map but did not apply it" and "needed heavy hints to find the bottleneck" show what a costly hint note looks like.
- Hollow acceptance. "Interviewers have seen candidates nod, say 'interesting,' and continue doing exactly what they were doing." Check whether your next sentence stated what the hint changes.
- "Let's move on" read as uniformly bad. Interviewer accounts conflict: it can mean the signal is already collected, or that you were stuck. Check whether you had stated a working approach before the cut.
- Early end read as rejection. By itself, it is neutral. Check whether you were invited to ask questions.
- Defending against a planted test. Some hints exist "to test whether you can evaluate input from a teammate." Accepting everything is also a weak signal. Check whether you gave a reason, either way.
- "What else?" answered with a random list. It feels productive but scores poorly. Check that each idea mapped to a named bucket.
- The felt-good round. One candidate described a round as their best by far, and it came back with mixed feedback. Your felt sense is not the scorecard.
These failure modes share one cause. The candidate is watching the interviewer's face for a verdict instead of producing evidence for the axis being tested.
The damaging note is never that you got a hint. It is that you could not use one.
The in-the-moment procedure, step by step
This is the order to run for any move, compressed so it fits inside the few seconds you have. Steps 1 to 3 apply to every move. Steps 4 to 8 apply to the matching move type. Step 9 happens after the round.
Responding to an interviewer move
- Classify the moveIn about two seconds, decide whether it is a hint, redirect, challenge, probe, time cue, or wrap. Done means you can name the type before you speak.
- Acknowledge and restateSpend about ten seconds repeating the hint or prompt back in your own words. Done means the interviewer has heard you restate it accurately.
- Integrate out loudAct on the hint within a minute while narrating what it changes. Done means your code, plan, or structure visibly changes.
- Ask a narrow question if stuckIf no hint is coming and you are past your preset threshold, ask one narrow question. Done means it states what you know, where you are blocked, and a proposed direction.
- Answer S/A/R probes directlyAnswer the follow-up first, without retelling the story. Done means the missing element, usually Action or Result, is supplied with specifics.
- Defend or update on challengesOn "are you sure?", run a trace or a test before answering. Done means you defend with that evidence or update and say why.
- Extend "what else?" in bucketsAdd new ideas inside the structure you already laid out. Done means every new idea sits in a named category.
- Close with prepared questionsUse the wrap for your prepared questions and do not ask how you did. Done means you asked them and noted what was said about next steps.
- Log the round within the hourWrite down concrete events separately from how the round felt. Done means every hint, redirect, probe, and challenge has a line with what you did next.
Step 2 is the one candidates skip under pressure, and it does double duty. It confirms you understood the move, and it gives the interviewer something to write down before your work even changes.
Who is across the table, and keeping the log current
More often than candidates expect, the person running the round interviews for a living. That makes a standardized hint ladder and a fixed probe set more likely. In Refolk's index of professional profiles, the title data points that way.
| Measure | United States | India |
|---|---|---|
| Interviewer-title profiles | 17 | 10 |
| At Karat | 2 | 2 |
| Karat share (derived) | 11.8% | 20% |
In India, Intervue.io and Codingal also appear among employers. Separately, 49 US profiles in Refolk's index list "Structured Interviewing" as a skill, with titles such as VP Talent Acquisition and Director of People Strategies, and Apple among the employers. Both samples are small, so read them as directional only. They suggest that a meaningful share of rounds are run by people trained to deliver the same moves to every candidate. That makes the moves more predictable, but it does not make the scorecard visible.
The post-round log
The log is what makes this reference cumulative. Each round's record tells you which moves keep landing on you, and that is the only honest read on a round you will get before the debrief.
Round: [company, round type, interviewer role if known] Scheduled length / actual length: [__ min / __ min] Moves (one line each): [minute] | [move type] | [what they said] | [what I did next] Hints: [asked or offered] | [what the hint was] | [minutes to apply it] Redirects: [had I stated a working approach before? yes/no] Probes: [which story] | [missing element: Situation, Action, or Result] Challenges: [held or updated] | [evidence I used] "What else?": [buckets named] | [did new ideas stay inside them?] Close: [invited to ask questions? yes/no] | [next steps, word for word] Feelings (separate): [how it felt]
Fill it within the hour; keep the Feelings line at the bottom and separate from everything above it.
Over several rounds, look for repeats. The same S/A/R gap twice means one story needs rewriting. Two "significant hint" rounds mean the gap is in preparation, not in-room technique. When the gap is in how a story is written down, the fix belongs on the page. Refolk's fit score for each posting also shows which requirements you match weakest, which is a reasonable guess at where the probes will dig next time.
Before you close the log
- Every hint has a line saying whether it was asked for or offered, and how many minutes it took to apply
- Every redirect has a yes or no on whether a working approach was already stated
- Every challenge records whether I held or updated, and the evidence I used
- Every behavioral probe names the missing element and the story it hit
- Scheduled and actual round length are both recorded
- Anything said about next steps is written word for word
- How the round felt sits on its own line, separate from the events
- I did not ask the interviewer how I did
Re-check the reference when your round format changes. Pair-programming rounds, case rounds, and structured behavioral panels run different hint and probe conventions. When the stakes justify it, check an interviewer-training source directly, such as the OPM guide or Google's re:Work page, rather than trusting any one prep vendor's rubric, including the ones quoted here.
Questions job seekers ask
Do hints count against you in a coding interview?
A hint you use quickly usually costs less than candidates fear. In public rubric reconstructions, "got there with hints" or "solution worked with hints" sits in the Lean Hire band, while "needed significant help" sits in Lean No Hire. The damaging note is a hint you could not apply. One self-described interviewer put it as hints are not a deal breaker, spoon feeding is. The bar is higher for senior candidates, where finding the path independently is what the round tests.
Is it a bad sign when the interviewer says let's move on?
Not by itself. Interviewer accounts conflict: one uses "I got the signal, let's move on" when a candidate is stuck, while others use it once the evidence is already collected. The useful check is whether you had already stated a working approach before the cut. Reply with a one-sentence summary of where you were and what was left, then move on without arguing. Log what you had stated before the move.
How long should I wait before asking for a hint?
Set your threshold before the round, because published advice disagrees by about 3x. Sources range from 3 minutes stuck, through 3 to 5 minutes of visible effort, to 8 to 10 minutes of real effort, and one treats asking outright as a last resort, no more than one or two times. They agree on the form: show your thinking first, then ask a narrow question that names where you are blocked and a proposed direction.
What should I do when the interviewer pushes back on my answer?
Test before you respond. Run a trace, check an example, or name the evidence, then either defend your answer with that evidence or update it and say what changed your mind. Rubric descriptions reward disagreeing with evidence and mark "defensive when challenged" as a weak signal. Folding instantly is weak too, because some pushback exists to test whether you can evaluate input from a teammate.
Does an interview ending early mean I failed?
No published outcome data supports that. One prep vendor claims ending about 30% early is a strong negative sign but gives no methodology, while an interviewer-side account reads early endings as a sign of doing well or doing exceptionally badly. By itself it is neutral. Check whether you were invited to ask questions, use the time for your prepared ones, and log how many minutes early it ended.
Should I ask the interviewer how I did at the end?
No. Final decisions are usually made by a panel afterward, and at most mid-to-large tech companies each interviewer writes an independent evaluation before a calibration debrief, typically within 24 to 72 hours of the final interview. The interviewer in front of you often cannot answer honestly yet. Use the closing minutes for prepared questions and write down, word for word, anything they say about next steps.
Put this to work
Paste your career in once. Every application after that is written for you.
Drop a resume or a LinkedIn URL. I rank the live openings against it, rewrite the resume and write a cover letter for the best of them, and fill in the employer's form when you press the button. You read, you decide what goes out.
01Drop your resume
A PDF or a LinkedIn URL. About a minute, once.
02I rank the openings
Every weekday morning, the live catalog scored against your history. Up to 20 worth your time, not two hundred links.
03Each one is written up
Resume rewritten for the posting, a cover letter, a fit score. Press send, or let me fill in the form.
- New matches ranked and written before you are up.
- Every bullet stays inside what your history supports. Nothing invented.
- Queued, submitted, interviewing, offer: one screen, not a spreadsheet.
Free, no card. Up to 25 applications sent a week, and nothing is sent until you say so.