Amazon Behavioral Interview Questions by Leadership Principle

Last updated (2026-11-02)

These Amazon behavioral interview questions are grouped by Amazon's 16 Leadership Principles, with a sample STAR answer for each one. They are typical examples of the kind of question asked for each principle, written for software engineers, new grads and interns applying to SDE roles in the US and worldwide.

The sample answers use fictional engineers. Treat them as a shape to follow, then replace every detail with a real story of your own.

What Amazon says about its behavioral interviews

Amazon publishes a fair amount about how it interviews software engineers. Everything in this section comes from Amazon's own pages; our Amazon interview process guide covers the stages in order and lists the sources.

  • The loop mixes question types: the SDE interview loop combines coding, system design (for SDE II) and behavioral questions based on the Leadership Principles.
  • Two or three behavioral questions per interviewer: Amazon's SDE II interview prep page says each interviewer typically asks two or three behavioral questions about past successes and challenges.
  • STAR is the recommended format: Amazon recommends answering with the STAR method (situation, task, action, result).
  • Every loop has a Bar Raiser: one interviewer, usually from outside the hiring team, is the Bar Raiser.
  • The online assessment touches the principles too: it includes a work style survey that also reflects the Leadership Principles.

In practice this means behavioral questions are not a short add-on at the end. Expect them in most interviews of the loop, alongside the technical questions.

Amazon's 16 Leadership Principles at a glance

Source: Leadership Principles on amazon.jobs, last checked 2026-11-02. The names below are Amazon's; the one-line summaries are ours, so read Amazon's page for the full wording.

Leadership PrincipleIn short (our summary)
Customer ObsessionBegin with the customer and work backwards from their needs, earn and keep their trust, and watch competitors while focusing on customers.
OwnershipThink long term, act on behalf of the whole company and not only your own team, and never treat something as not your job.
Invent and SimplifyExpect innovation from yourself and others, keep simplifying, take good ideas from anywhere, and accept being misunderstood for a long time when doing something new.
Are Right, A LotShow strong judgment and good instincts, look for diverse perspectives, and actively try to prove your own beliefs wrong.
Learn and Be CuriousKeep learning and improving, stay curious about new possibilities, and act on that curiosity by exploring them.
Hire and Develop the BestRaise the bar with every hire and promotion, spot talent, develop leaders and take coaching others seriously.
Insist on the Highest StandardsHold standards that others may find unreasonably high, keep raising them, do not pass defects along, and fix problems so they stay fixed.
Think BigSet and communicate a bold direction, think differently, and look ahead for new ways to serve customers.
Bias for ActionMove fast, because many decisions can be reversed and do not need long study, and value calculated risks.
FrugalityDo more with less, treat constraints as a source of resourcefulness, and earn no credit for headcount, budget size or fixed costs.
Earn TrustListen closely, speak candidly, treat people with respect, criticize yourself openly, and compare yourself and your team with the best.
Dive DeepWork at every level, stay close to the details, check often, question it when metrics and anecdotes disagree, and treat no task as beneath you.
Have Backbone; Disagree and CommitRespectfully challenge decisions you disagree with even when it is uncomfortable, do not give in just to keep the peace, and fully commit once a decision is made.
Deliver ResultsFocus on the inputs that matter, deliver them with the right quality and on time, push through setbacks and do not settle.
Strive to be Earth's Best EmployerWork toward a safer, more productive, higher performing, more diverse and more just workplace, lead with empathy, and care whether colleagues are growing and empowered.
Success and Scale Bring Broad ResponsibilityBe humble and thoughtful about the secondary effects of your work on communities, the planet and future generations, and leave things better than you found them.
The questions on this page are typical examples of the kind of question asked for each principle, written by us. They are not leaked or official Amazon questions, and your interviewers may phrase things very differently.

How to answer Amazon behavioral interview questions

Our behavioral interview questions guide explains the STAR method and how to build a story bank in general. For an Amazon loop, a few points matter more than usual.

  • Make Action the longest part. Situation and Task are context; keep them to a sentence or two each. Spend most of your time on what you actually did and why.
  • Say "I", not "we". Team context is fine, but the interviewer is assessing your decisions. If you only describe what the team did, expect to be asked what your part was.
  • End with a result you can describe. A measurable result (latency, error rate, time saved) is strong when you have one. An observable result (the bug stopped recurring, the team adopted your checklist) is fine too. Never invent numbers.
  • Expect follow-ups. Interviewers often dig into one story with several follow-up questions about details, tradeoffs and what you would do differently. Only tell stories you know well enough to go two or three levels deeper.
  • Failures are fair game. Questions about mistakes, missed deadlines and wrong calls are common. Pick a real one, own your part, and spend time on what changed afterwards.
  • Reuse stories carefully. Each interviewer is usually assessing different principles, so telling the same story to everyone wastes a chance to show range. Reusing one story across two interviewers, told from a different angle, is less risky than a weak story.
  • New grads have plenty of material. Internships, class team projects, hackathons, open source contributions, research and part-time jobs all count. Interviewers mostly care about your reasoning, not the size of the project.

Prepare stories that each cover several principles, then decide which angle to lead with based on the question. For example:

  • An on-call incident you debugged: Dive Deep (finding the root cause), Ownership (staying with it), Insist on the Highest Standards (the fix that stopped it recurring).
  • A design you pushed back on: Have Backbone; Disagree and Commit, Earn Trust, Are Right, A Lot.
  • A feature shipped under a tight deadline: Deliver Results, Bias for Action, Frugality.
  • A new tool or language you picked up for a project: Learn and Be Curious, Invent and Simplify.
  • Helping a teammate or new intern get productive: Hire and Develop the Best, Strive to be Earth's Best Employer.
Practice each story out loud at least twice, ideally with someone asking follow-ups. The AI mock interview lets you practice role-based questions out loud and get feedback.

Customer Obsession interview questions

Typical examples: for an engineer, this usually means a story where a user's real problem, not the spec or your own preference, decided what you built.

1.Tell me about a time you went beyond what was asked to solve a customer's problem.

What they check: whether you think about the person using the product, including internal customers, and not only about closing your ticket.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: I owned a report export API used by our support team. A support lead filed a ticket saying exports for large accounts timed out.

Task: The ticket only asked me to raise the timeout, which would have taken ten minutes.

Action: Before changing it, I sat with two support agents for half an hour to watch how they used the export. They were exporting full histories just to find the last week of activity. I raised the timeout as a stopgap, then added a date range filter and made it the default in their tool.

Result: Typical exports dropped from about two minutes to a few seconds, and the support lead told me the team stopped batching exports overnight.

Follow-ups to expect: How did you know the timeout was not the real problem? What would you have done if the date filter had taken two weeks?

2.Describe a time you had to balance what a customer asked for with what was technically right.

What they check: that you work backwards from the customer's underlying need, and can say no to a specific request while still solving the problem.

Sample answer (STAR)Fictional new-grad engineer.

Situation: In my senior capstone, we built a scheduling app for a campus tutoring center. The center's director asked us to email every tutor whenever any session changed.

Task: I owned notifications, and I worried that tutors would start ignoring a flood of emails.

Action: I asked the director what problem the emails were solving. It turned out tutors were missing last-minute cancellations. I proposed emailing only the affected tutor, plus a daily summary, and showed her a mockup of both before building anything.

Result: She agreed, and in the two weeks of pilot use no tutor missed a cancelled session, which was the original problem.

Ownership interview questions

Typical examples: strong Ownership stories show you fixing something outside your formal scope, or thinking about the long-term cost of a shortcut.

3.Tell me about a time you took on something outside your area of responsibility.

What they check: whether you act when something is broken even if it is not technically yours, and whether you do it in a way that respects the real owner.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: Our team's deploys kept failing on a shared integration test suite owned by another team. Everyone retried the pipeline and moved on.

Task: Nobody had been asked to fix it, but it was costing my team several hours a week.

Action: I tracked failures for a week and found three tests that depended on wall clock time. I messaged the owning team with the data, offered to send the fix, and opened a pull request that injected a fake clock. Their engineer reviewed and merged it.

Result: Retries on our pipeline dropped to near zero over the next month, and the other team added a lint rule so new tests could not read the real clock.

Follow-ups to expect: Why not just disable the tests? What would you have done if the other team had not responded?

4.Give an example of a time you made a decision with long-term impact in mind rather than a short-term win.

What they check: that you weigh future maintenance and the wider team, not just the fastest path to done.

Sample answer (STAR)Fictional new-grad engineer.

Situation: During my internship, I added a new field to a customer profile page. The quickest fix was to copy the same validation logic into three different forms.

Task: I had one sprint left, and the copy approach would have fit easily.

Action: I asked my mentor whether a shared validator was worth it. We agreed it was, so I pulled the logic into one module with tests and updated all three forms, and I wrote a short note in the team wiki on where validation now lived.

Result: I finished a day before the sprint ended. A month later, my mentor told me a follow-up change took one line in the shared module instead of three separate edits.

Invent and Simplify interview questions

Typical examples: good stories here show you replacing a complicated process with a simpler one, or trying a new approach others doubted.

5.Tell me about a time you simplified a complex process or system.

What they check: whether you notice unnecessary complexity and remove it, rather than adding another layer on top.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: Setting up a local development environment on my team took new engineers most of their first two days, following a 40-step wiki page.

Task: I was onboarding a new teammate and wanted to cut that time down.

Action: I wrote a single script that installed dependencies, seeded a local database and ran a health check, and replaced the wiki page with three commands. I asked the new teammate to use it first and fixed the two steps that broke for her.

Result: The next two hires had a working environment within their first morning, and the wiki page went from 40 steps to a short troubleshooting list.

6.Describe a new idea you proposed that others were skeptical of at first.

What they check: that you can invent something useful and stick with it while others are unconvinced, using evidence instead of insistence.

Sample answer (STAR)Fictional new-grad engineer.

Situation: At a 36-hour hackathon, my team planned a mobile app for finding study rooms on campus. I suggested skipping the app and using the room booking system's existing calendar feed instead.

Task: My teammates thought a web page reading a calendar feed would look less impressive.

Action: I built a rough prototype in two hours that showed free rooms on a map from the live feed, and showed it to them. We agreed to spend the saved time on a better filter for quiet rooms and accessibility.

Result: We had a working demo with real data by the deadline and placed second, and two judges specifically mentioned that it worked with live data.

Are Right, A Lot interview questions

Typical examples: this principle shows up in stories about decisions made with incomplete data, and about changing your mind when evidence pointed the other way.

7.Tell me about a time you made a good decision with incomplete information.

What they check: how you reason under uncertainty: what you knew, what you assumed, and how you limited the risk of being wrong.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: We had to pick a message queue for a new event pipeline, and the expected traffic was a rough estimate from the product team.

Task: I was asked to recommend an option within a week, without load test data.

Action: I listed what we knew for sure (ordering per user mattered, exactly-once did not) and what was a guess (peak volume). I chose the managed option our team already operated, because it handled our known needs and could scale further than the high estimate, and I wrote down the traffic level at which we should revisit.

Result: Six months later, peak traffic was close to the middle estimate and the pipeline had no capacity incidents. The revisit threshold in the design doc was referenced when traffic grew.

Follow-ups to expect: What would have made you choose differently? How did you decide what was a safe assumption?

8.Tell me about a time you changed your mind after hearing a different perspective.

What they check: that good judgment includes actively seeking out views that disagree with yours.

Sample answer (STAR)Fictional new-grad engineer.

Situation: In a databases course project, I was sure we should store user activity in a document database because the schema felt flexible.

Task: As the team's backend lead, the final call on the data model was mine.

Action: Before deciding, I asked a teammate who preferred a relational design to write down the queries our reports needed. Most of them were joins and aggregations across users and courses, which would have been awkward with my design.

Result: I switched to a relational schema. Our reporting features took about half the time we had planned, and I now list the real queries before choosing a data store.

Learn and Be Curious interview questions

Typical examples: engineers usually answer these with a story about picking up an unfamiliar technology quickly, or digging into how something works out of curiosity.

9.Tell me about a time you had to learn a new technology quickly to get something done.

What they check: how you learn: what resources you use, how you check your understanding, and how fast you become useful.

Sample answer (STAR)Fictional new-grad engineer.

Situation: At the start of my internship, I was assigned a bug in a service written in Go. I had only written Java and Python before.

Task: My mentor expected a fix within the first two weeks.

Action: I spent the first two evenings on the official Go tour, then read the service's tests before its code, because they showed how each function was meant to be used. I kept a list of questions and asked my mentor in one fifteen-minute block each day instead of interrupting constantly.

Result: I shipped the fix in the second week, and later in the internship I wrote a small Go endpoint on my own.

10.What is something you learned recently on your own, and how have you used it?

What they check: genuine curiosity outside assigned work, and whether you turn learning into action.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: I kept seeing flame graphs mentioned in performance discussions but had never used one.

Task: Nobody asked me to, but our main API had a slow endpoint I wanted to understand.

Action: I read a few profiling guides over a weekend, then profiled the endpoint in our staging environment. The flame graph showed most time going to JSON serialization of a large nested object that the client never used.

Result: I proposed trimming the response, the client team agreed, and the endpoint's median latency dropped by roughly a third. I then gave a 20-minute demo on profiling to my team.

Hire and Develop the Best interview questions

Typical examples: you do not need to have hired anyone; stories about mentoring an intern, onboarding a teammate or giving useful feedback fit well.

11.Tell me about a time you helped a teammate grow or improve.

What they check: whether you invest in other people's growth, and how you do it without just doing their work for them.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: A summer intern on my team was struggling with code reviews: his pull requests got dozens of comments and he seemed discouraged.

Task: I was his onboarding buddy, so helping him was partly my job.

Action: I suggested he send me small draft changes before opening a pull request, and in our weekly chat we went through the most common review comments together. I also showed him how to split a large change into smaller ones.

Result: By the end of the summer his pull requests were smaller and needed far fewer review rounds, and he presented his project to the wider team on his own.

12.Describe a time you gave difficult feedback to a peer.

What they check: whether you raise the bar for the people around you, and whether your feedback is specific and kind.

Sample answer (STAR)Fictional new-grad engineer.

Situation: In a software engineering course project, one teammate regularly merged code without tests, which broke the build for the rest of us.

Task: I was the informal team lead and did not want to escalate to the professor.

Action: I asked him for a quick call and described two specific broken builds and how much time they cost, without blaming him. I asked whether anything was making testing hard for him. He said he did not know how to mock the database, so I paired with him for an hour on it.

Result: His next pull requests came with tests, the build stayed green for the rest of the semester, and he later helped another teammate with the same thing.

Insist on the Highest Standards interview questions

Typical examples: strong stories show you refusing to ship something not good enough, or fixing a problem at its root so it stays fixed.

13.Tell me about a time you refused to compromise on quality.

What they check: that you hold a quality bar under pressure, and that you can explain why it was worth it.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: We were close to launching a redesigned checkout form. During my own testing with a screen reader, I found that error messages were not announced at all.

Task: The launch was two days away and my manager asked whether it could wait for a follow-up.

Action: I explained that users relying on screen readers would not be able to complete checkout at all, which made it a blocker rather than a polish item. I fixed the error announcements and focus order the same day, and added an automated accessibility check to our pipeline for that page.

Result: We launched on the original date, and the new check caught two further accessibility regressions in later releases.

Follow-ups to expect: How did you decide this was a blocker? What would you have done if the fix had taken a week?

14.Describe a time you fixed a problem so that it would not happen again.

What they check: whether you stop at the symptom or go on to the root cause and a lasting fix.

Sample answer (STAR)Fictional new-grad engineer.

Situation: During my internship, a nightly data job failed about once a week, and the team's habit was to rerun it by hand the next morning.

Task: I was asked to rerun it when it failed, but I wanted to know why it failed.

Action: I compared logs from failed and successful runs and found the job started before an upstream export had finished. I changed it to wait for the export's completion marker, and added an alert if the export was more than an hour late.

Result: The job did not fail again during the remaining eight weeks of my internship, and nobody had to rerun it by hand.

Think Big interview questions

Typical examples: an engineer's Think Big story usually shows you looking past the immediate task toward a larger direction, and persuading others to follow it.

15.Tell me about a time you proposed a bigger vision than what was originally planned.

What they check: whether you can see a larger opportunity and communicate it clearly enough that others back it.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: I was asked to add retry logic to one service that called a flaky partner API.

Task: While working on it, I noticed four other services had their own slightly different retry code for the same partner.

Action: I wrote a one-page proposal for a shared client library with retries, timeouts and metrics, showing the four existing copies side by side. I offered to build the first version and migrate my service first as a proof.

Result: My manager approved it. Three of the four services moved to the library within the quarter, and we got one dashboard for partner errors instead of none.

16.Describe a project where you set an ambitious goal. How did you approach it?

What they check: that you aim high but break the goal into realistic steps, and are honest about what you achieved.

Sample answer (STAR)Fictional new-grad engineer.

Situation: I maintained a small open source library for parsing course timetables used by students at a few universities.

Task: I set myself a goal of supporting every university format people had requested in the issue tracker, not just my own.

Action: I designed a plugin system so each format could be added separately, wrote a guide for contributors, and built two plugins myself as examples.

Result: Over the next semester, contributors added four more formats. I did not reach every request, but the plugin design meant I no longer had to write each one myself.

Bias for Action interview questions

Typical examples: these stories show you moving quickly on a reversible decision instead of waiting for perfect information, while managing the risk.

17.Tell me about a time you had to act quickly without all the information you wanted.

What they check: that you can tell reversible decisions from irreversible ones and act fast on the first kind.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: While on call, I got an alert that error rates on our search API had jumped right after a deploy, but the logs did not yet show an obvious cause.

Task: I had to decide whether to keep investigating or roll back.

Action: Rolling back was quick and safe because the deploy had no database changes, so I rolled back first and confirmed error rates recovered. Then I reproduced the issue in staging and found a missing config value for a new feature flag.

Result: Customer-facing errors lasted about ten minutes. We added a startup check for required config, and I redeployed the fixed version the next morning.

Follow-ups to expect: When would you not roll back first? What did you communicate, and to whom, during the incident?

18.Give an example of a calculated risk you took.

What they check: that you take risks deliberately, with a plan for limiting the damage if you are wrong.

Sample answer (STAR)Fictional new-grad engineer.

Situation: My intern project was a new caching layer for a read-heavy internal API. My mentor was away for a week right when it was ready.

Task: I could wait a week or launch it carefully without him.

Action: I put the cache behind a feature flag, turned it on for 5 percent of traffic, and watched error rates and latency for a day before increasing it. I wrote down the exact steps to turn it off and shared them with the team.

Result: The rollout reached all traffic by the end of the week with no incidents, and my mentor came back to a working feature instead of a waiting one.

Frugality interview questions

Typical examples: engineers often answer with a story about cutting cloud costs, reusing existing tools, or delivering with a smaller team or budget than expected.

19.Tell me about a time you accomplished something with limited resources.

What they check: that constraints made you more resourceful rather than stalling the work.

Sample answer (STAR)Fictional new-grad engineer.

Situation: My student club wanted an events site, but we had no budget for hosting or paid tools.

Task: I volunteered to build and run it.

Action: I built a static site generated from a shared spreadsheet that club officers already used, hosted it on a free static hosting tier, and used a free form service for sign-ups. I documented how officers could add events without touching code.

Result: The site ran all year at no cost, and officers added events themselves without needing me.

20.Describe a time you found a way to reduce costs without hurting quality.

What they check: whether you notice waste and fix it, while protecting what customers rely on.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: Our team's monthly cloud bill had grown steadily, and nobody had looked at where it went.

Task: I offered to spend a few days finding savings.

Action: I broke the bill down by service and found that test environments ran around the clock and old log data was kept in the most expensive storage tier. I scheduled test environments to shut down outside working hours and moved logs older than 30 days to cheaper storage, after confirming with the team how often they read old logs.

Result: The monthly bill dropped by about a fifth, and nobody reported missing anything. I set up a monthly cost report so we would notice future growth early.

Earn Trust interview questions

Typical examples: Earn Trust stories often involve admitting a mistake openly, or rebuilding a working relationship through candor and respect.

21.Tell me about a time you made a mistake. How did you handle it?

What they check: whether you are vocally self-critical: you own the mistake quickly and fix the process, not just the symptom.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: I wrote a database migration that added an index on a large table. I tested it on a small staging dataset, but in production it locked the table for several minutes.

Task: Some requests failed during that window, and the incident was caused by my change.

Action: I posted in the incident channel that the migration was mine and what I believed had happened. After recovery, I wrote the incident review, and I proposed that index migrations on large tables use the database's non-locking option and run against a production-sized copy first.

Result: The team adopted both changes as part of our migration checklist, and we have not had a migration lock since.

Follow-ups to expect: What did you say in the incident channel? What would you do differently today?

22.Describe a time you had to rebuild trust with a teammate or stakeholder.

What they check: that you listen, speak candidly and treat people with respect, especially when things have gone wrong.

Sample answer (STAR)Fictional new-grad engineer.

Situation: In a group project, I rewrote a teammate's module overnight without asking because I thought it was too slow. She was upset when she found out.

Task: I needed to repair that relationship, since we had six weeks of work left together.

Action: I apologized directly and said I should have raised my concern with her first. I reverted my rewrite, shared the benchmark that worried me, and asked how she wanted to handle it. We improved her version together.

Result: The final module was faster than both versions, and for the rest of the project we reviewed each other's changes before merging. She asked to team up with me again the next semester.

Dive Deep interview questions

Typical examples: this is often where engineers tell debugging stories, especially ones where the numbers and the user reports did not agree.

23.Tell me about the most difficult bug you have debugged.

What they check: how you dig through layers to a root cause, and whether you can explain technical detail clearly.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: One of our integration tests failed roughly once in every fifty runs, and nobody could reproduce it locally.

Task: The flaky test blocked deploys often enough that I was asked to find the cause.

Action: I ran the test in a loop with extra logging and saw that failures happened only when two test cases ran in parallel. Both wrote to the same temporary file path. I traced the path to a helper that generated names from the current second, so parallel tests occasionally collided.

Result: I switched the helper to unique temporary directories. The test has not failed since, and I found and fixed two other tests using the same helper.

Follow-ups to expect: How did you narrow it down to parallel runs? Why was it not reproducible locally? What did you rule out first?

24.Describe a time the data said one thing but your instinct or customer feedback said another.

What they check: that you stay skeptical when metrics and anecdotes differ, and investigate instead of trusting the dashboard by default.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: Our dashboard showed page load times were fine, but several customers reported the app felt slow every morning.

Task: I owned frontend performance, so I took the reports.

Action: I checked how the metric was calculated and found it was an average across the whole day, which hid a morning spike. Looking at hourly percentiles, the slowest loads were three times worse right after a nightly cache flush.

Result: We changed the flush to warm the cache afterward, morning load times matched the rest of the day, and we switched the dashboard to hourly percentiles.

Have Backbone; Disagree and Commit interview questions

Typical examples: good answers show you disagreeing respectfully with evidence, and then fully committing once a decision is made, even if it went against you.

25.Tell me about a time you disagreed with your manager or team lead.

What they check: that you raise disagreement openly and respectfully, backed by data, rather than staying quiet or complaining later.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: My manager wanted to skip load testing for a new notifications feature to hit a launch date.

Task: I was the lead engineer on the feature and thought skipping it was risky because it fanned out to every user at once.

Action: I raised it in our one-on-one with a rough estimate of peak messages per second compared with what the downstream service had handled before. I proposed a two-day load test limited to the fan-out path rather than the full suite.

Result: My manager agreed to the smaller test. It found a connection pool limit we fixed before launch, and we still shipped on the original date.

Follow-ups to expect: What would you have done if your manager had said no? How did you make sure it did not feel like a personal challenge?

26.Describe a time you committed to a decision you did not agree with.

What they check: the second half of the principle: once a decision is made, you support it fully instead of undermining it.

Sample answer (STAR)Fictional new-grad engineer.

Situation: In my internship team, I argued for writing a new service in Python because I knew it well. The team chose Java to match their other services.

Task: I was going to write most of the service, in a language I was less comfortable with.

Action: I accepted the decision and did not reopen it. I spent my first week going through the team's existing Java services to follow their conventions, and asked for early reviews so I would not drift from them.

Result: The service shipped on schedule, and my mentor said it was easy to maintain because it looked like the rest of their code. Looking back, the team's reasoning about consistency was right.

Deliver Results interview questions

Typical examples: these stories show you getting something important done on time and at the right quality despite setbacks along the way.

27.Tell me about a time you delivered a project under a tight deadline.

What they check: how you prioritize the key inputs, cut scope sensibly and still deliver with the right quality.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: A partner integration had to go live in three weeks because of a contract date, and my original estimate was five weeks.

Task: I was the only engineer on it and needed a plan that fit.

Action: I listed every feature and asked the product manager which ones the partner needed on day one. We cut two reporting features to a follow-up and kept full test coverage on the payment path. I posted a short status every two days so surprises surfaced early.

Result: We went live a day early with no launch issues, and the reporting features shipped two weeks later.

28.Describe a time you faced a major setback but still delivered.

What they check: that you rise to the occasion when things go wrong, and do not settle for an excuse.

Sample answer (STAR)Fictional new-grad engineer.

Situation: Two weeks before our capstone demo, the free API our app depended on for transit data shut down.

Task: I was responsible for the data layer, so finding a replacement fell to me.

Action: I found that the city published the same schedules as open data files. I wrote a parser for them and a small nightly job to refresh the data, and kept the same interface so the rest of the app did not change.

Result: We demoed on time with real schedules. The professor noted the app no longer depended on a third-party service that could disappear.

Strive to be Earth's Best Employer interview questions

Typical examples: engineers can answer these with stories about making a team more inclusive, safer or easier to grow in, even without being a manager.

29.Tell me about a time you made your team a better place to work.

What they check: whether you care about your colleagues' experience and growth, and act on it.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: Our on-call rotation was stressful: the same few people handled most pages because others did not know the systems well.

Task: I was one of the people taking most of the pages, and I wanted the load shared fairly without throwing newer teammates in at the deep end.

Action: I wrote runbooks for our five most common alerts and set up a shadow rotation where newer engineers paired with an experienced one for their first two on-call weeks.

Result: Within two months everyone on the team was on the primary rotation, and in our team survey on-call stress came up far less often than before.

30.Describe a time you helped someone who was struggling to feel included or supported.

What they check: empathy and whether you notice when someone is being left out.

Sample answer (STAR)Fictional new-grad engineer.

Situation: On a hackathon team, one member who was new to programming rarely spoke and was left with no clear task.

Task: I was coordinating the team and noticed she seemed unsure where to start.

Action: I asked her privately what she wanted to learn. She was interested in design, so I asked her to own the user flow and paired with her on the frontend for an hour so she could build part of it herself.

Result: She built the onboarding screens and presented them during our demo, and she told me later it was the first hackathon where she felt part of the team.

Success and Scale Bring Broad Responsibility interview questions

Typical examples: engineers can answer these with stories about thinking through the side effects of their work on users, communities or the environment.

31.Tell me about a time you considered the wider impact of a technical decision.

What they check: whether you look beyond your immediate goal to the secondary effects of what you build.

Sample answer (STAR)Fictional SDE with two years of experience.

Situation: We planned to add detailed analytics events to our mobile app to understand feature usage.

Task: I was implementing the event tracking.

Action: While reviewing the event list, I noticed several events would have captured free-text search queries, which could contain personal information. I raised it with the product manager and privacy reviewer, and proposed logging only the search category and result count instead.

Result: The team adopted the change, and we still got the usage insight we needed without storing what users typed.

32.Describe something you did to leave a project, codebase or community better than you found it.

What they check: that you take responsibility for the long-term state of things you touch, beyond your own task.

Sample answer (STAR)Fictional new-grad engineer.

Situation: I contributed a small feature to an open source accessibility checker. While setting it up, I found the contributor guide was out of date and several new contributors had given up in the issue tracker.

Task: My feature was done, and nobody asked me to fix the guide.

Action: I rewrote the setup section, tested it on a fresh machine, and labeled five simple issues as good first issues with short notes on where to start.

Result: The maintainers merged the new guide, and over the next two months several first-time contributors picked up the labeled issues.

Put this guide to work

Build a one-page resume, check it against a real job description, then practice your answers out loud.

Frequently asked questions

How many leadership principles does Amazon have?

Amazon's Leadership Principles page lists 16 principles, from Customer Obsession to Success and Scale Bring Broad Responsibility. Behavioral questions in the SDE loop are based on them.

Are these real Amazon interview questions?

No. They are typical examples of the kind of question asked for each Leadership Principle, written by us. They are not leaked or official Amazon questions, so prepare flexible stories rather than answers to exact wording.

How many stories do I need for Amazon leadership principles interview questions?

Many candidates prepare around eight to ten strong stories and map each one to two or three principles. That gives enough range to avoid repeating the same story to every interviewer in the loop.

Should I memorize Amazon behavioral interview questions and answers?

Memorize the key facts of your own stories, not scripted answers. Interviewers ask follow-ups that go deeper into details, and a memorized answer tends to break down under that questioning.

Can a new grad use school projects for Amazon behavioral questions?

Yes. Internships, class team projects, hackathons, open source contributions and part-time jobs all work. What matters is that the story is real, that you explain your own actions clearly, and that you can answer follow-ups about it.