What this prompt does
This prompt turns a blank page into a structured coaching session for behavioral interviews. You feed it your target company, its specific values or principles, the questions you want to practice, and your experience level — and it returns a complete answer kit: not just a framework, but a working strong example, a deliberately weak example with a diagnosis, and the follow-up question the interviewer is likely to fire next.
What makes it actually useful is the contrast structure. Seeing a weak answer dissected alongside a strong one trains your instincts faster than reading advice alone. The template also asks for red flags interviewers watch for, which surfaces the anti-patterns most candidates never hear about until after they fail.
The story banking section at the end is the part I find most underused. It pushes you to map your existing experiences to multiple question types rather than preparing 12 separate stories — a much more sustainable prep strategy for senior roles where the question set is unpredictable.
When to use it
- You have a behavioral round scheduled at Amazon, Google, Meta, or a company with documented values, and want answers aligned to their specific rubrics.
- You are a senior or staff engineer prepping for leadership-heavy behavioral loops where generic STAR advice is too shallow.
- You want to stress-test existing answers by generating the weak version and checking whether yours avoids those failure modes.
- You are coaching a junior teammate and need structured teaching examples, not just bullet points.
- You are switching industries and need to reframe technical achievements into leadership or cross-functional impact language.
- You have limited prep time and want a prioritized list of company-specific questions beyond the obvious ones.
Example output
Question: "Tell me about a time you disagreed with a technical decision."
Tests: Amazon LP — Have Backbone; Disagree and Commit
STAR Template:
Situation: What was the decision context and your role?
Task: What was at stake if the decision went unchallenged?
Action: How did you raise the disagreement — to whom, with what evidence?
Result: What was the outcome? Did the team change course or did you commit?
Strong answer (senior engineer level):
"In Q3 we were about to migrate our payments service to an async queue
architecture. I believed the latency tail risk was underestimated. I wrote
a one-page doc with p99 data from our staging load tests, brought it to
the tech lead 48 hours before sign-off, and proposed a bounded sync
fallback path. The team adopted the fallback. In production, it triggered
3 times in the first month and prevented one P1."
Weak answer:
"I told my manager I disagreed and they eventually came around."
Why it fails: No evidence of how you disagreed. No data. No demonstration
that you understood the other side before pushing back.
Red flags: Vague outcomes, blaming teammates, no mention of listening.
Likely follow-up: "What would you have done if the team still disagreed after
your doc?"
Pro tips
- Set
[experience_level]honestly. A staff engineer answer that leans on "I learned a lot" reads as mid-level and will cost you. The prompt calibrates the example to your level — trust it. - Use
[count]: 1on your first pass. Work one question to completion, record your own answer, then compare. Running five questions at once tempts you to skim. - For Amazon specifically, set
[company_values]to the exact LP name (e.g., "Ownership, Dive Deep") rather than a summary. The prompt uses that text when explaining which principle a question probes — precision matters. - Pair this with a second pass where you set
[questions]to the follow-up questions this prompt generated. Most candidates prepare openers; interviewers remember how you handle the second-level question. - After generating your story bank, run the prompt once more with
[target_company]swapped to a backup company. Good stories transfer, but the framing of what "success" looks like shifts — and the prompt will surface that difference.
