Applied Scientist
Amazon Leadership Principles (Behavioral)
Amazon has the most structured and intensive behavioral interview process in tech. Every question maps to one of 16 Leadership Principles. You will have dedicated LP rounds โ often 2+ hours of purely behavioral questions. The Bar Raiser is a senior Amazonian who checks that you meet the LP bar. Prepare 2 strong STAR stories per LP.
Core Theory
- 1The 16 Leadership Principles: (1) Customer Obsession โ leaders start with the customer and work backwards. (2) Ownership โ think long-term, act like an owner, never say 'that's not my job'. (3) Invent and Simplify โ innovate and find ways to simplify. (4) Are Right, A Lot โ strong judgment, seek diverse perspectives. (5) Learn and Be Curious โ never stop learning. (6) Hire and Develop the Best โ raise the bar with every hire. (7) Insist on the Highest Standards โ never settle. (8) Think Big โ bold direction. (9) Bias for Action โ speed matters, calculated risks are good. (10) Frugality โ accomplish more with less. (11) Earn Trust โ candor, humility, transparency. (12) Dive Deep โ stay connected to details. (13) Have Backbone; Disagree and Commit โ challenge decisions respectfully, then commit fully. (14) Deliver Results โ focus on the key inputs, deliver with the right quality. (15) Strive to be Earth's Best Employer. (16) Success and Scale Bring Broad Responsibility.
- 2Bar Raiser role: a trained Amazonian (not on the hiring team) who ensures candidates meet the company-wide LP bar, not just the team's immediate need. They vote independently and can override the hiring team. They will probe 3 levels deep into your stories: 'What did YOU specifically do? Why did you choose that approach? What would you do differently?' Prepare for this depth.
- 3STAR format at Amazon: Situation (1-2 sentences โ set minimal context), Task (your specific role and what was at stake), Action (most important โ concrete, first-person, detailed, explain your reasoning), Result (quantified impact + long-term effects). Amazon interviewers specifically note if you use 'we' too much โ they want to know what YOU did.
- 4Dive Deep probing pattern: interviewers will ask follow-up questions to verify you actually did what you claim. 'Tell me more about how you identified that root cause.' 'What was the exact data that showed you this?' 'How did you specifically convince your manager?' Prepare to answer 3 layers deep on any story.
- 5Story selection strategy: each story should cover 2-3 LPs. Prioritize: Customer Obsession (always asked), Ownership (often asked twice), Dive Deep (technical roles), Deliver Results (always), Earn Trust (especially for senior roles). Have 6 distinct stories to avoid repeating the same one when the same LP comes up in multiple rounds.
- 6Customer Obsession at Amazon: this is THE most important LP. Every decision traces back to the customer. Even internal tools must be thought of as serving customers. Story ingredients: customer pain point discovered, action taken to understand the customer's perspective deeply (wrote a press release, used the product yourself, interviewed customers), solution implemented, customer metric improved.
- 7Disagree and Commit: shows backbone without being combative. Structure: disagreement should be based on data or principles, not preference. You raised it professionally with evidence. The decision went against you. You committed fully and worked to make it successful. You may or may not have been vindicated โ either outcome can be a good story.
- 8Avoid anti-patterns: (1) 'I can't think of a failure' โ every strong candidate has a genuine failure. (2) 'We decided to...' instead of 'I decided to...' (3) Stories without measurable results. (4) Blaming others for the failure. (5) Stories where you were the hero who was right and everyone else was wrong โ Amazon values humility and collaboration.
Key Formulas
| 1 | # โโโ Amazon LP Story Bank Template โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ |
| 2 | |
| 3 | LP_STORIES = { |
| 4 | "Customer Obsession": { |
| 5 | "questions": [ |
| 6 | "Tell me about a time you went above and beyond for a customer.", |
| 7 | "Describe a time you had to make a tradeoff between a customer need and a business constraint.", |
| 8 | "Tell me about a time you saw a customer need that others hadn't noticed.", |
| 9 | ], |
| 10 | "story_template": """ |
| 11 | SITUATION: [Customer problem discovered โ how you found out about it] |
| 12 | TASK: [Your role, what was at stake for the customer] |
| 13 | ACTION: |
| 14 | - Specifically: HOW did I discover the customer problem? (data, user interviews, support tickets) |
| 15 | - What data did I gather to understand the problem deeply? |
| 16 | - What specific actions did I take, and why those specifically? |
| 17 | - What tradeoffs did I navigate to prioritize the customer? |
| 18 | RESULT: |
| 19 | - Customer metric improved: [NPS +10, support tickets -40%, CTR +8%] |
| 20 | - What did the customer experience change to? |
| 21 | """, |
| 22 | "strong_signals": [ |
| 23 | "Went beyond your scope to solve a customer problem", |
| 24 | "Used data to understand the customer (not just assumed)", |
| 25 | "Made a short-term cost tradeoff to benefit the customer long-term", |
| 26 | ] |
| 27 | }, |
| 28 | |
| 29 | "Ownership": { |
| 30 | "questions": [ |
| 31 | "Tell me about a time you took ownership of a problem outside your scope.", |
| 32 | "Tell me about a time you saw something going wrong and fixed it proactively.", |
| 33 | "Describe a time you had to make a decision without guidance.", |
| 34 | ], |
| 35 | "story_template": """ |
| 36 | SITUATION: [Problem that was 'nobody's job' or a crisis without a clear owner] |
| 37 | TASK: [Why you stepped in, what was at risk] |
| 38 | ACTION: |
| 39 | - What specifically made me take ownership when I could have ignored it? |
| 40 | - What decisions did I make independently? |
| 41 | - What barriers did I overcome? |
| 42 | RESULT: |
| 43 | - What would have happened without my intervention? |
| 44 | - Quantified impact of my actions |
| 45 | """, |
| 46 | }, |
| 47 | |
| 48 | "Dive Deep": { |
| 49 | "questions": [ |
| 50 | "Tell me about a time you found the root cause of a problem that others had missed.", |
| 51 | "Describe a time you caught a significant error by staying close to the details.", |
| 52 | "Tell me about a time you had to analyze a large dataset to make a decision.", |
| 53 | ], |
| 54 | "story_template": """ |
| 55 | SITUATION: [Problem that appeared simple but had a non-obvious root cause] |
| 56 | TASK: [Why deep investigation was necessary, stakes involved] |
| 57 | ACTION: |
| 58 | - What data did I dig into? (be specific about tables, metrics, queries) |
| 59 | - What intermediate hypotheses did I test and discard? |
| 60 | - What was the insight that others had missed? |
| 61 | - How did I verify I had the right root cause? |
| 62 | RESULT: |
| 63 | - What was the actual root cause vs. the apparent one? |
| 64 | - Impact of finding the real root cause |
| 65 | """, |
| 66 | }, |
| 67 | |
| 68 | "Disagree and Commit": { |
| 69 | "questions": [ |
| 70 | "Tell me about a time you disagreed with your manager and how you handled it.", |
| 71 | "Describe a time you were overruled. How did you respond?", |
| 72 | "Tell me about a time you challenged a decision with data.", |
| 73 | ], |
| 74 | "story_template": """ |
| 75 | SITUATION: [Decision being made that you disagreed with] |
| 76 | TASK: [Your role, what was at stake] |
| 77 | ACTION: |
| 78 | - What specifically was I disagreeing with? (be precise โ not just 'I had a bad feeling') |
| 79 | - What data or principle did my disagreement rest on? |
| 80 | - How did I raise my concern? (who, how, what evidence) |
| 81 | - How did I commit after the decision was made? |
| 82 | - What did I do to make the decision work even though I disagreed? |
| 83 | RESULT: |
| 84 | - Outcome โ was I right or wrong? (both can be good stories) |
| 85 | - What did this accomplish for the team/organization? |
| 86 | """, |
| 87 | }, |
| 88 | } |
| 89 | |
| 90 | # โโโ Bar Raiser Probe Patterns โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ |
| 91 | BAR_RAISER_PROBES = [ |
| 92 | "You said 'we analyzed the data' โ what specifically did YOU analyze?", |
| 93 | "You said you 'convinced' your team โ walk me through exactly how you did that.", |
| 94 | "Why did you choose THAT approach specifically? What alternatives did you consider?", |
| 95 | "How did you know the root cause was X? What data confirmed it?", |
| 96 | "What was the customer's reaction when you delivered the result?", |
| 97 | "What would you do differently if you could do it again?", |
| 98 | "How do you know this was YOUR impact vs. other factors?", |
| 99 | ] |
| 100 | |
| 101 | print("LP Story Framework loaded.") |
| 102 | print(f"LPs covered: {list(LP_STORIES.keys())}") |
| 103 | print(f"\nBar Raiser probes to prepare for:") |
| 104 | for probe in BAR_RAISER_PROBES: |
| 105 | print(f" โข {probe}") |
A structured framework for Amazon LP story preparation. The story templates show exactly what details each LP question is looking for. The Bar Raiser probe bank shows the follow-up questions you must be able to answer โ these are the questions that separate candidates who 'have stories' from candidates who 'lived the stories.'
Worked Interview Problems
3 problemsProblem
Prepare a strong Customer Obsession story for an Applied Scientist role at Amazon.
Solution
Situation: 'I was reviewing our recommendation model's performance dashboard when I noticed that one specific user segment โ new parents โ had dramatically lower engagement than the average. This wasn't tracked by any of our standard KPIs.'
Task: 'I wasn't asked to investigate this. My OKR was to improve overall CTR. But I believed that improving this segment could matter both to customers and to the business long-term.'
Action: 'I pulled the raw events for this segment and found they were receiving product recommendations that made no sense for their likely life stage โ no baby products, no family content, heavy gaming recommendations. I traced this to a cold-start model that didn't use demographic signals. I spent 3 days (outside my normal sprint) building a prototype that incorporated new-parent signals from their purchase history. I tested it on a 5% holdout and showed +18% CTR for the segment. I then wrote a 2-page memo to leadership with the data and proposed it as a Q3 initiative.'
Result: 'The initiative was approved. I led the implementation over 6 weeks. Engagement for new parents improved 22%, and our post-survey showed +14 NPS improvement from this group. This became the foundation of our life-stage recommendation framework, which I presented at the team's annual review as a key technical achievement.'
Answer
Strong signals: customer pain discovered proactively (not assigned), went outside scope, used data to understand root cause, quantified customer impact (+22% engagement, +14 NPS), created lasting organizational impact (life-stage framework).
Common Mistakes
Using 'we' throughout without specifying your individual contribution. Amazon is measuring your leadership, not your team's. Every sentence in the Action should be 'I did X because Y.'
Picking failure stories that are secretly humble brags: 'I worked too hard and burned out,' 'I was too detail-oriented.' These are red flags. Choose a real failure with real consequences and genuine learning.
Stories without data or quantification. 'The customer was happier' is weak. 'Customer NPS improved from 42 to 56, and support tickets for this feature dropped 67%' is strong. If you don't have exact numbers, estimate: 'approximately $1M annually based on the volume.'
Not being ready to go 3 levels deep. A Bar Raiser will ask: 'You said you analyzed the data โ what exact query or analysis did you run?' If you can't answer, the story loses credibility. Prepare the technical details.
Choosing stories where the outcome was entirely positive and you were the hero. Amazon values intellectual humility. Stories where you were wrong, learned something, and changed your behavior often score higher than stories where you were right all along.