A practice prompt we wrote. No company or candidate report names it, so it carries no company tag.
How to answer
Some companies weigh this answer openly. As of September 2026, Anthropic’s US application form asks “Why do you want to work at Anthropic?”, says it values the response highly and notes that great answers are often 200-400 words. Source 1Forward Deployed Engineer application formPublisherAnthropic (Greenhouse job board)Source typecompany job posting A Palantir blog post from October 2025 says Palantir screens in interviews for mission commitment by looking for evidence of it in candidates’ backstories. Source 2Beyond Founder Mode: Mission Mode (Palantir Blog)PublisherPalantir BlogSource typecompany blog In a Palantir video from July 2025 aimed at software engineering candidates, a hiring manager told them not to give a prepared statement of what they think Palantir wants to hear. Source 3From Code to Career: Meet Palantir's Hiring Managers (YouTube uploaded English captions)PublisherPalantir, YouTubeSource typerecorded talk or interview That is not a reason to wing it. Prepare evidence, not a script: the line you quote, the thing you built and the decision you made. Then say them in your own words each time.
The test for every sentence is the same: swap in a competitor’s name. If the sentence still works, it isn’t an answer to this question. Three kinds of sentence pass it:
- Something they published, quoted. A line from the posting, an engineering post or a product decision with a date on it, and where you read it. “Your posting says an engagement ends when the customer’s own team runs the system” can be said to one company; “I love your mission” can be said to all of them.
- Something you did with their product. What you built with it, or read about how customers deploy it, and the part that was harder than the docs suggested. You would be deploying it, so this sentence is yours alone.
- A decision of yours that matches. Not a belief but a choice with a cost: the job you left, the project you pushed for, the shortcut you refused. This is the backstory evidence Palantir’s blog describes.
Try the swap test on three weak sentences. Each one fails because a competitor’s name fits; each fix names something only this company published, or something only you did.
| Fails the swap test | Passes it |
|---|---|
| “I love your mission of making AI useful.” | “Your posting says an engagement ends when the customer’s own team runs it.” |
| “You’re the leader in the space.” | “I ran your connector on a public dataset; the permissions model was the hard part.” |
| “I’m passionate about customers.” | “I stayed through a customer’s peak season on a fixed fee so their own team could run the system.” |
“Why now” has two halves. The company’s half is its own dated FDE decision, such as a team launch or its oldest live FDE posting; Why the role spread when it did shows how to find it. Your half is what you just finished, or what your current role can no longer give you, said as a pull toward this team rather than an escape. A market-size figure answers neither half and fails the swap test. Close with one sentence on what you would bring to their first engagements.
The follow-ups check the reading. For “What else did you read from us, and what did you disagree with?”, have one respectful disagreement ready, such as a trade-off in their deployment model you would have made differently, and why. For “Which other companies are you talking to?”, name them if you are comfortable doing so, then give the one thing this company has that they don’t, which should be the published line you opened with. For “What would make you leave in your first year?”, name a condition about the work, such as losing customer contact after the first engagement, not pay or perks.
In a recruiter screen, say it in about a minute. A written answer on an application form can carry the full version. Write it separately for each target, and reread what the company has published in the week before the interview. The answer below is the written version, about the length Anthropic’s form suggests for its own “why” question. Source 1Forward Deployed Engineer application formPublisherAnthropic (Greenhouse job board)Source typecompany job posting The spoken cut follows it and keeps only the posting line, the decision, why now and the close.
Researching a company and turning what you find into this answer is taught in Behavioral mastery for FDEs, in Pro. The free lesson on reading an FDE job posting line by line shows how to find the sentence worth quoting.
Follow-ups
What the interviewer may ask next, once your first answer is on the table.
- What else did you read from us, and what did you disagree with?
- Which other companies are you talking to, and why would you pick us?
- What would make you leave in your first year?
Where answers go wrong
- Praises the mission in the company’s own adjectives and could say the same sentence to any competitor.
- Answers “why now” with a market statistic instead of something about this company and the candidate’s own timing.
Answer this in two minutes
Write the answer you would say out loud. The clock starts with your first word.
Compare with the model answer
Illustrative answer about a fictional project
Written in the first person to show the structure. Tell yours from your own work.
What drew me first is a line in your posting: that an engagement ends when the customer’s own team runs the system, not when the pilot works. I also read the post your platform team wrote in May about moving each customer’s custom code into a repository the customer owns, so that a deployment survives after your engineers leave.
I also ran your connector against a public freight dataset last weekend. The schema mapping took an hour, and the part I’d expect to be hard at a customer is the permissions model, which is where our handovers usually break.
That matches the decision I’m proudest of. At my current company I build integrations for logistics customers. Last year we had a pilot with a cold-chain distributor that went well, and the plan was to hand it to their IT team with a document. I argued that we should stay through the first peak season and hand over a runbook and alerting they had practiced with, not a document. It cost us six extra weeks on a fixed-fee contract, and my manager pushed back. We stayed. In the second week of peak, their temperature feed dropped out at one warehouse at 3 a.m., and their on-call engineer fixed it from the runbook in 20 minutes without calling us. That is what made them renew, and it is the work your posting describes.
Why now: your posting says this is the team’s first year with dedicated FDEs outside the home office and that the early hires will write the deployment playbook. I’ve written that playbook once, informally, at a company where deployment supports the product. I’d like to write it where deployment is the product. On my side, we just finished moving our last three large customers onto a partner-run integration model. That was the right call for my company, and it means the work I want to do there is ending.
What I’d bring to your first engagements is the habit of planning the handover from the first week: who on the customer side will own it, what they need to practice, and what we stay for.
Said in a minute:
Your FDE posting says an engagement ends when the customer’s own team runs the system, not when the pilot works. That’s the work I want. Last year a cold-chain pilot of ours was meant to end with a handover document. I argued that we should stay through their peak season and hand over a runbook their team had practiced, which cost us six extra weeks on a fixed fee. In the second week of peak, a temperature feed dropped out at 3 a.m., and their on-call engineer fixed it from the runbook without calling us. Why now: your team is in its first year with FDEs outside the home office, and the early hires will write the playbook. I’ve written one informally where deployment supports the product; I want to write it where deployment is the product. I’d bring the habit of planning the handover from the first week.