A practice prompt we wrote. No company or candidate report names it, so it carries no company tag.
How to answer
The interviewer wants to know whether you care more about the result than the thing you made. Colin Jarvis, who leads forward deployed engineering at OpenAI, said on a South Park Commons panel posted in May 2026 that a failure pattern is people who come to love the form of what they built more than its function, and that the best FDEs tear a solution up and build something different when users do not use it. Source 1The Future of Forward Deployed Engineering | OpenAI, Ramp, Nominal, DatalandPublisherSouth Park Commons (YouTube)Source typerecorded talk or interview
Five moves, with the most time on the second and third:
- What you built, and why it was a fair bet. Two sentences, with “I” as the subject. If it was obviously wrong from the start, the story is about judgment, not attachment.
- The evidence. What users did instead of using it: went back to the spreadsheet, exported to email, called someone. Say where you saw it (logs, a sit-along, a support queue) and give a count, not a feeling: how many of the intended users opened it after the first week. Then the rescue you tried, and the deadline and bar you set for it before you started.
- The call. Who you told, starting with whoever asked for it or reported it upward as a win, in what words, what the build had cost, and what you kept: the data pipeline, the test set, the access you had negotiated. Then how you turned it off: access removed, users told, nothing left running just in case.
- What replaced it, often smaller and plainer than what you threw away, and its result.
- What you do earlier now. The check that would have told you sooner.
Say the kill in terms of the customer’s goal: “The planners weren’t opening it, and the goal was fewer missed orders, not a new screen.” It shows where your attachment sits.
The first trap is a story with nothing to give up: a spike built to be discarded, or a project your manager canceled. The second is blaming the users: what they did instead is your best evidence, so describe it.
Follow-ups
What the interviewer may ask next, once your first answer is on the table.
- What did you try before you decided to kill it, and how long did you give that attempt?
- Who objected to throwing it away, and what did you tell them?
- What from the discarded build did you keep, and where is it now?
Where answers go wrong
- Picks a prototype that was always meant to be thrown away, or a project someone else canceled, so there was no attachment to overcome and no call of their own.
- Blames the users for not adopting it, or gives the evidence as a feeling, instead of naming what users did instead.
Answer this in two minutes
Write the answer you would say out loud. The clock starts with your first word.
Illustrative answer about a fictional project
I was the deployed engineer for a group of 22 veterinary clinics, at a company that sold inventory analytics. I built a reorder dashboard for the clinic managers.