Capstone: Mock Loop
Run the full gauntlet - self-exam, drills and a readiness check
Timed scheduling-tetris drill
After this you can build and fix a loop under time pressure.
You don't find out whether your scheduling holds up by reading about buffers and time zones. You find out by starting a 10-minute timer, opening a blank calendar and building a clean loop while narrating every choice. This section is that drill, built on the exact shape Cursor's practical round would hand you: a panel, a candidate flying in and constraints that don't quite fit.
Set the timer before you read the constraints. The clock is part of the test. The practical exercise is checking whether you can hold five moving pieces in your head and still produce something a candidate could follow without a single follow-up question.
Interactive diagram. Tab through its regions; each focused region shows its detail in the panel below.
This capstone drills the exact stages Cursor's RC loop puts you through - step through each to see what it tests.
- Candidate
- Priya, currently in San Francisco (PT), flying to the Soho NYC office for an in-person loop on Thursday.
- Loop
- 4 interviews + a 45-min lunch host slot, all between 10:00 and 16:00 ET on Thursday.
- Panel
- Recruiter (ET), Hiring manager (ET), two cross-functional interviewers - one ET, one PT joining over Zoom.
- Hard constraints
- The PT interviewer is unavailable before 12:30 ET; the HM has a standing block 13:00–14:00 ET.
- Goal
- A back-to-back loop with buffers, an explicit itinerary and complete context routed to every interviewer.
Build the loop - 10-minute clocknarrate as you place each block
- 1Convert to one zone first. Pin everything to ET, the office zone and translate the PT interviewer's 12:30 ET floor (9:30 their time) so you never quote a wrong number. Write the candidate's home zone next to her name so the invite shows both.
- 2Place the hard constraints before the easy ones. The PT interviewer can only land in the afternoon and the HM is blocked 13:00–14:00 - solve those two against each other first, then fill recruiter and lunch around the gaps.
- 3Add buffers, don't pack to the minute. Leave 10–15 minutes between interviews for the candidate to breathe and for a slow interviewer to wrap; bake a real lunch in, not a working one she eats alone.
- 4Assign rooms and Zoom explicitly. In-person slots get a room name; the PT interviewer's slot gets a Zoom link in the invite and on the itinerary so nobody hunts for it at 14:00.
- 5Write the candidate itinerary last. One clean schedule in her home zone and ET both, with names, roles, format (in-person vs Zoom) and where to physically be.
- ET slot
- 10:00–10:45
- Interview
- Recruiter sync
- Who / zone
- Recruiter (ET)
- Format
- In-person, Room A
- ET slot
- 11:00–12:00
- Interview
- Hiring manager
- Who / zone
- HM (ET)
- Format
- In-person, Room A
- ET slot
- 12:00–12:45
- Interview
- Lunch host
- Who / zone
- You + candidate
- Format
- In-person
- ET slot
- 12:45–13:45
- Interview
- Cross-functional
- Who / zone
- PT interviewer (9:45 PT)
- Format
- Zoom, link in invite
- ET slot
- 14:00–15:00
- Interview
- Cross-functional
- Who / zone
- ET interviewer
- Format
- In-person, Room A
| ET slot | Interview | Who / zone | Format |
|---|---|---|---|
| 10:00–10:45 | Recruiter sync | Recruiter (ET) | In-person, Room A |
| 11:00–12:00 | Hiring manager | HM (ET) | In-person, Room A |
| 12:00–12:45 | Lunch host | You + candidate | In-person |
| 12:45–13:45 | Cross-functional | PT interviewer (9:45 PT) | Zoom, link in invite |
| 14:00–15:00 | Cross-functional | ET interviewer | In-person, Room A |
PT floor (12:30 ET) and the HM 13:00–14:00 block both respected; lunch placed where it naturally falls.
Mid-drill, the ET cross-functional interviewer at 14:00 drops with a production incident. Don't freeze and don't silently rebook. Surface 2–3 options fast: (1) swap in a pre-cleared backup interviewer for the same 14:00 slot, (2) move the 12:45 Zoom interviewer to 14:00 and pull the recruiter sync earlier to keep the loop back-to-back or (3) if no backup exists, hold the four confirmed and offer the candidate a short Zoom finish the next morning rather than stretching her travel day. State the tradeoff of each in one sentence and recommend one.
Narrate your prioritization out loud the whole time, exactly as you would in the live exercise: “I'm placing the PT interviewer and the HM block first because they're the only fixed points, then everything flexible fills around them.” The screener is grading your reasoning as much as the final grid. Silent correctness reads as luck; narrated correctness reads as judgment under load.
Self-check before you call it done
- Every time is stated in both zones and you triple-checked the PT translation.
- Buffers exist between every interview and lunch is a real break.
- Every interviewer has complete context: candidate name, role, resume, interview kit, what to assess.
- The candidate has one unambiguous itinerary: where to be, who she's meeting, in-person vs Zoom and your phone number for day-of.
- Axis
- Accuracy
- Weak (1-2)
- Quoted a slot in the wrong zone; missed a hard constraint.
- Strong (4-5)
- All constraints respected; both zones correct; buffers present.
- Axis
- Speed
- Weak (1-2)
- Still rearranging at the buzzer.
- Strong (4-5)
- Clean loop placed in time, with room to absorb the curveball.
- Axis
- Communication
- Weak (1-2)
- Built the grid in silence.
- Strong (4-5)
- Narrated prioritization; itinerary readable by a stranger.
- Axis
- Recovery
- Weak (1-2)
- Panicked or rebooked one option without options.
- Strong (4-5)
- Offered 2–3 recovery paths with tradeoffs and a recommendation.
| Axis | Weak (1-2) | Strong (4-5) |
|---|---|---|
| Accuracy | Quoted a slot in the wrong zone; missed a hard constraint. | All constraints respected; both zones correct; buffers present. |
| Speed | Still rearranging at the buzzer. | Clean loop placed in time, with room to absorb the curveball. |
| Communication | Built the grid in silence. | Narrated prioritization; itinerary readable by a stranger. |
| Recovery | Panicked or rebooked one option without options. | Offered 2–3 recovery paths with tradeoffs and a recommendation. |
Anything you score 3 or below is a rep to repeat before the real exercise.
Takeaway. Pin everything to the office zone, place hard constraints first, leave real buffers and when an interviewer drops, surface 2–3 recovery options with tradeoffs instead of silently rebooking.
Self-check
QMid-loop, a back-to-back interviewer drops 30 minutes before their slot. What's the strongest first move?
Candidate-comms writing test
After this you can write the four core candidate messages to a high bar.
Coordination is mostly writing. The candidate who flew across the country judges Cursor partly by your emails, so the comms test grades whether your words are warm, exact and impossible to misread. Draft all four core messages, then grade them against the rubric below.
The four messages every loop needs: an interview invite, a prep note, a status update and a rejection. Write each one, then go back and cut anything that reads like template boilerplate. Cursor wants comms that feel written by a person, not pasted from a system.
Interactive diagram. Tab through its regions; each focused region shows its detail in the panel below.
The candidate hears from you at each step - the rejection is the quality gate that most reveals your judgment.
1. Interview invite - timezone-explicit
Subject: Your Cursor onsite - Thursday 6/19, Soho NYC Hi Priya, We'd love to host you at our Soho office this Thursday, June 19. Your loop runs 10:00 AM–3:00 PM ET (7:00 AM–12:00 PM PT) - I'll send a full itinerary with names and rooms once we lock it. Address: [office], buzz suite 4. Ask for me at the desk and I'll come down to meet you. Lunch is on us between sessions. Could you confirm this works by EOD tomorrow? If anything about travel or timing is tricky, reply here and we'll flex around you. Looking forward to meeting you, [Name] · [phone]
2. Prep note - set them up to win
Subject: Prep for Thursday - what to expect Hi Priya, A few notes so Thursday feels easy: - You'll meet four people: [HM] (hiring manager), [recruiter] and two team members - one in person, one over Zoom (link in your invite). - One session is a practical coordination scenario; it's conversational, not a quiz. Think out loud and you'll be in great shape. - Dress is casual. Bring whatever helps you - notes are fine. - My cell is below for anything day-of, including running late. You've got this. See you Thursday, [Name] · [phone]
3. Status update - short, honest, no dead air
Subject: Quick update on next steps Hi Priya, Wanted to keep you in the loop: the team is wrapping up debriefs and I expect to have clear next steps for you by Tuesday. Nothing for you to do right now - I just didn't want you sitting in silence. Thanks for your patience and reach out anytime, [Name]
4. Rejection - warm, human, leaves the door open
Subject: An update on your Cursor application Hi Priya, Thank you for the time you gave us - flying in and meeting the team is no small thing and we genuinely enjoyed having you in the office. After the debrief, we've decided not to move forward for this role. This was a close, considered decision. If you're open to it, I'd love to stay in touch as the team grows. Wishing you the best and thank you again, [Name]
Grade every draft against this rubric
- Dimension
- Warmth
- Fails the bar
- Transactional, system-generated tone.
- Meets the bar
- Sounds like a person who's glad they're here.
- Dimension
- Clarity
- Fails the bar
- Time without a zone; vague next step.
- Meets the bar
- Both zones; one obvious next action and deadline.
- Dimension
- Personalization
- Fails the bar
- Generic, could be any candidate.
- Meets the bar
- Names the role, the people, the specifics of their day.
- Dimension
- Zero ambiguity
- Fails the bar
- “We'll be in touch soon.”
- Meets the bar
- “Clear next steps by Tuesday” - a date, not a vibe.
| Dimension | Fails the bar | Meets the bar |
|---|---|---|
| Warmth | Transactional, system-generated tone. | Sounds like a person who's glad they're here. |
| Clarity | Time without a zone; vague next step. | Both zones; one obvious next action and deadline. |
| Personalization | Generic, could be any candidate. | Names the role, the people, the specifics of their day. |
| Zero ambiguity | “We'll be in touch soon.” | “Clear next steps by Tuesday” - a date, not a vibe. |
Phrases that scream boilerplate: “per my last email,” “please be advised,” “we regret to inform you,” and any time quoted without a zone. The rejection is the hardest one and the most revealing: a cold rejection costs Cursor a referral and a future re-application. Read each draft aloud - if it sounds like a form letter, rewrite the opening line first.
If the comms test is take-home, send your drafts back with a one-line note on your choices: “I led the rejection with gratitude and kept the door open because a candidate who flew in deserves that and re-applications are real.” If it's live, have a peer or the rubric review your tone before you submit. Showing your reasoning about tone is itself a signal of the judgment the role needs.
Takeaway. Every candidate message should be warm, both-zones explicit and personalized, with one obvious next action - and the rejection deserves the most care, not the least.
Self-check
Ashby / ops mini-exercise
After this you can demonstrate ATS hygiene and a scoped improvement idea.
Hiring managers make decisions off whatever the ATS shows them. If the pipeline is messy, they're deciding on bad data and the loop slows down while everyone reconciles. This exercise hands you a tangled Ashby pipeline and grades whether you can clean it and propose one improvement worth making.
The mess in front of you: candidates sitting in the wrong stage, completed interviews with no scorecard attached and an interview kit pointing the wrong panel at the wrong round. Fix it in order, then propose exactly one process change - scoped tight enough that you could ship it this week.
Triage the messy pipeline in order
- Mis-staged candidates first - anyone who's actually further along than their stage shows. A candidate parked in “Screen” who already did the onsite blocks the HM's view of the real shortlist; move them to true stage and note why.
- Missing scorecards next - find completed interviews with no submitted feedback and chase the interviewer, because a decision made without their input is the costliest gap.
- Broken interview kits last - confirm each stage routes the right panel, the right kit and the right scorecard template before the next candidate hits it.
- Then reconcile the record - names, stages and dates that disagree with the calendar or email thread get corrected to a single source of truth.
Propose one improvement - before / after
Scorecards trickle in days after interviews; HMs ping you to chase each one.
Debrief gets scheduled only once the last scorecard lands.
Median time from final interview to decision: ~5 days.
Scorecard reminder fires automatically 1 hour post-interview; a daily digest flags any still missing at 24h.
Debrief is pre-booked when the loop is built, not after feedback lands.
Target: final interview to decision under 2 days.
Don't declare victory on a hunch. Check two things over the next handful of loops: did median time-to-decision actually drop and did scorecard completion-at-24h go up without interviewers complaining the reminders are noise. Watch for the side effect - an over-eager reminder that pings someone who already submitted. Confirm the automation respects “already done” before you scale it past one team.
- Trustworthy data for HMs
- A clean pipeline means the HM's shortlist view is real, so debriefs run on facts instead of “wait, where is this person?”
- Faster loops
- Cursor runs ~2-week loops; every day shaved off feedback-to-decision is a day the candidate isn't drifting to a competing offer.
- Less interviewer friction
- Automated nudges beat you personally chasing busy engineers, which protects the relationship you need for the next loop.
Show judgment about scope. The trap is proposing a giant ATS overhaul; the signal is picking the one change with the highest friction-removed-per-effort and saying why you'd leave the rest alone for now. “I'd ship the scorecard reminder first because it's one config change that fixes the slowest part of our loop. A full kit redesign can wait until we've felt the gain.” That restraint is the initiative-plus-judgment the role is built on.
Takeaway. Clean the pipeline in order - mis-staged candidates, missing scorecards, broken kits - then propose one tightly scoped improvement with a before/after metric and a plan to verify it didn't break anything.
Self-check
QYou've cleaned a messy pipeline and want to propose one process change. What makes a proposal land in this exercise?
Behavioral & values mock
After this you can rehearse the values screen end to end.
The values rounds at Cursor screen for warmth and ruthless reliability at the same time, plus genuine excitement about the mission. You can't fake either under pressure, so you rehearse: run the likely questions out loud, in STAR, each one mapped to a value the company actually holds.
Run these six prompts as a full mock. Answer each in STAR - situation, task, action, result - and quantify the result wherever you honestly can. Record yourself or grab a peer, then review the tape for the two signals these rounds chase: warmth and reliability.
Interactive diagram. Tab through its regions; each focused region shows its detail in the panel below.
Map every STAR story to one of these - warmth and reliability are co-equal at the top.
- 1The complex schedule. “Walk me through the hairiest loop you ever coordinated.” Map to detail obsession and calm under load. End on the outcome: candidate showed up on time across three zones, zero reschedules.
- 2The experience save. “Tell me about a time you turned a rough candidate moment into a good one.” Map to hospitality and empathy. Quantify if you can - they accepted or referred someone later.
- 3The conflict. “A busy interviewer keeps no-showing your invites.” Map to stakeholder management and reliability. Show how you set expectations without burning the relationship.
- 4Discretion. “You saw comp data you shouldn't share. What did you do?” Map to discretion and trustworthiness. Keep it crisp and unembellished - over-explaining here reads as the opposite of discreet.
- 5Initiative. “What did you change that nobody asked you to?” Map to ownership and initiative. Tie it to a before/after, like the scorecard reminder from the ops exercise.
- 6Failure. “Tell me about a coordination mistake.” Map to truth-seeking. Own it plainly, say what you changed and don't hide behind a humblebrag.
Specific situation, your actions (not the team's), a quantified result.
Named the value without forcing it: “this was a reliability thing for me.”
Honest about the messy parts; warmth comes through unprompted.
Vague, hypothetical or all “we” with no “I.”
No result or a result you can't quantify or even describe.
A failure story that's secretly a brag, which fails the truth-seeking screen.
The 90-second “why Cursor, why this role”
“I want to operate the hiring machine for a company that's actually shipping the future of how code gets written. Cursor runs fast in-person loops with fly-in candidates and a multi-day onsite project and that only works if coordination is flawless and the candidate feels genuinely hosted. That's the exact intersection I'm good at - I triple-check the timezone and I love making someone's hard travel day feel easy. The pace and the mission are the draw, not the deterrent.”
Time it. If it runs past 90 seconds you're rambling; if it's under 30 you're underselling. The structure that lands: mission, why this role specifically and one concrete thing about how you work that proves you mean it.
Closing questions and a confident close
- Ask something that shows you've thought about the actual job: “What's the part of the loop that breaks most often when you're hiring fast and how would you want me to own it?”
- Ask about the team, not just the perks: “How does the team make sure a fly-in candidate leaves feeling great even when it's a no?”
- Close with intent: thank them by name, say plainly you want the role and confirm next steps and timing before you leave the room.
Review your recording for warmth and reliability separately - they're different signals and most people are strong on one and thin on the other. If you sound warm but vague on specifics, you'll read as nice-but-unreliable. If you're precise but flat, you'll read as competent-but-cold. The role demands both at once, so rerun any answer that only lands one of them.
Takeaway. Answer every behavioral prompt in STAR mapped to a real Cursor value, quantify results, keep “why Cursor” under 90 seconds and review your tape for warmth and reliability as separate signals.
Self-check
QAn interviewer asks: “Tell me about a coordination mistake you made.” What does a strong answer do and what's the trap?
Readiness self-assessment
After this you can score your readiness and target your last-mile prep.
The last mile isn't more cramming. It's an honest scan of where you're sharp and where you're soft, so the hours before the loop go to the reds, not to re-reading what you already own. Rate each area red, yellow or green - and be ruthless, because a generous self-grade is the most expensive shortcut here.
Treat the rubric as a gap finder, not a score to admire. Mark the weakest proof, turn it into one practice rep and keep the artifact you would show an interviewer.
- Area
- Scheduling craft
- Green means
- You build a multi-zone loop with buffers and recover from a drop in under 10 min.
- If yellow/red, do this
- Re-run the section 1 drill twice with a fresh curveball each time.
- Area
- ATS / ops
- Green means
- You can clean a messy pipeline and pitch one scoped fix with a metric.
- If yellow/red, do this
- Redo the section 3 triage; write the before/after from memory.
- Area
- Candidate experience
- Green means
- Your four core messages are warm, both-zones, personalized.
- If yellow/red, do this
- Rewrite the rejection until it sounds human, not templated.
- Area
- Behavioral / values
- Green means
- Six STAR stories, each mapped to a value and quantified.
- If yellow/red, do this
- Record the weakest two answers; re-tape until warmth + reliability both land.
- Area
- Cursor knowledge
- Green means
- You can name the real process cold.
- If yellow/red, do this
- Drill the four facts below until they're reflexive.
| Area | Green means | If yellow/red, do this |
|---|---|---|
| Scheduling craft | You build a multi-zone loop with buffers and recover from a drop in under 10 min. | Re-run the section 1 drill twice with a fresh curveball each time. |
| ATS / ops | You can clean a messy pipeline and pitch one scoped fix with a metric. | Redo the section 3 triage; write the before/after from memory. |
| Candidate experience | Your four core messages are warm, both-zones, personalized. | Rewrite the rejection until it sounds human, not templated. |
| Behavioral / values | Six STAR stories, each mapped to a value and quantified. | Record the weakest two answers; re-tape until warmth + reliability both land. |
| Cursor knowledge | You can name the real process cold. | Drill the four facts below until they're reflexive. |
Spend your last-mile hours on rows you marked red, in this order of impact.
Name Cursor's actual process - cold
- Fast loops
- Cursor runs unusually short ~2-week interview loops; speed of coordination is a direct lever on whether elite talent says yes.
- No-AI screens
- Technical screens are run without AI assistance, a deliberate policy you should know and not be surprised by.
- Multi-day paid onsite
- The real decision round is a multi-day paid onsite project, which is the loop's centerpiece to host well.
- Soho in-person
- The job is in-person at the Soho NYC office; Cursor weights onsite presence and you'll host fly-in candidates there.
If any of these surprises you in the room, the values panel will notice.
Story bank coverage check
Walk your stories against the JD's pillars and Cursor's values. Every pillar should have at least one story; every value should have at least one. Gaps are where the interviewer will find air.
- Pillars: complex scheduling, candidate hosting, comms hub, stakeholder collaboration, ATS hygiene, process improvement, discretion.
- Values: truth-seeking, ownership, reliability, detail obsession, empathy, calm under load, discretion, hospitality.
- If a pillar or value has no story, build one before the loop - even a small, true one beats a confident blank.
Confirm the obvious things that the role is literally about: you replied to every recruiter email within hours, you triple-checked your own interview times and zones, you know exactly where the Soho office is and when to arrive and you've prepped warm, specific questions. The strongest possible signal is being the warm, reliable, prepared, on-time candidate you'd want greeting people at the door. Model the job before you have it.
Takeaway. Grade yourself honestly red/yellow/green, spend the last mile on the reds, have Cursor's four facts reflexive and walk into the loop as the warm, reliable, on-time candidate you'd hire.
Self-check
QName the four facts about Cursor's hiring process you should be able to state cold in a values round.