Where it comes from

How to answer

As of September 2026, the US application form asks “Why do you want to work at Anthropic?”, says the response is valued highly and notes that great answers are often 200-400 words. Source 1Forward Deployed Engineer application formPublisherAnthropic (Greenhouse job board)Source typecompany job posting Anthropic’s candidate AI guidance asks you to write the first draft of an application yourself, then use Claude to refine it. Source 3Guidance on Candidates' AI UsagePublisherAnthropicSource typecompany website Expect to defend it out loud.

The written answer has three parts.

  1. A principle you engage with, not recite. Anthropic’s published principles include “Put the mission first”, which calls the mission the final arbiter in its decisions. Source 2CareersPublisherAnthropicSource typecompany hiring page Engaging means saying what you think it could cost in your seat: a customer not getting what it asked for. Then give a decision of yours where you paid a similar cost.
  2. The work, from the posting. The FDE postings describe building production applications with Claude inside customer systems and delivering servers, sub-agents and skills. Source 1Forward Deployed Engineer application formPublisherAnthropic (Greenhouse job board)Source typecompany job posting Name the closest thing you have built and the part that was hard.
  3. Why now. The postings call these hires founding FDEs who help shape the forward-deployed motion. Source 1Forward Deployed Engineer application formPublisherAnthropic (Greenhouse job board)Source typecompany job posting Say what you have done early in something that others then built on.

The defense. The skeptical follow-up tests the first part, and “it’s written down” fails it, because a written principle is what the interviewer is doubting. Answer with a published decision that had a cost. In September 2026 Anthropic released Claude Fable 5.1 and Claude Mythos 5.1, the same model with different safeguards, and made Mythos 5.1 available only through trusted access programs. Source 4Claude Fable 5.1 and Mythos 5.1PublisherAnthropicSource typecompany website Say why that counts as evidence to you, say you can’t see how it was decided, and point back to your own decision. For the competitor question, say yes or no, then say what would decide it.

The trap is mission vocabulary with nothing of yours underneath it.

GlossaryForward deployed engineerA software engineer who builds and ships production systems inside a customer’s problem and environment, accountable to that customer’s outcome.More on Forward deployed engineerGlossaryModel Context ProtocolAn open protocol for exposing tools and data to models through servers, so agents can call a customer’s systems in a standard way.More on Model Context ProtocolGlossaryAgentA system in which a model chooses steps and tool calls to complete a task, within limits the design sets.More on Agent

Follow-ups

What the interviewer may ask next, once your first answer is on the table.

  • Every lab says it puts safety first. Why should I believe it here?
  • When would putting the mission first cost a customer something, and what would you do?
  • Would you take the same role at another lab?

Where answers go wrong

  • Repeats safety and mission language with no decision of the candidate’s own behind it, so the skeptical follow-up has nothing to push against.
  • Claims to know how decisions are made inside the company, from a forum post or a guess, and is corrected by someone who works there.

Answer this in two minutes

Write the answer you would say out loud. The clock starts with your first word.

Two minutes

Illustrative answer about a fictional project

Last year I recommended that a customer not let our model stop benefit payments on its own, and it cost us a quarter on the renewal. I lead integrations at a company that sells claims-processing software to state unemployment agencies.

Learn this in Module 11: Behavioral mastery for FDEs, part of the curriculum.