0 day streak

Amazon behavioral interview: Leadership Principles

Big techProductE-commerce and consumer US and India Bar raiser Machine coding / LLD

Each interviewer gets two to four principles to test and asks STAR questions about them, even in coding and design rounds. Interviewers write down quotes as evidence and argue hire or no-hire in a debrief run by the Bar Raiser.

0%ready
ProgressPrincipleThey will ask something likeYour story
Customer ObsessionCustomer focus Describe a time you changed a plan because of what you learned about the user.Tell me about a time you had to balance a customer need against a technical constraint. Write a story
Sample answer: Tell me about a time you went out of your way for a customer.

Start from what the customer needs and work backwards; keep their trust even when a competitor or an internal goal pulls you elsewhere.

Situation I worked on the sign-up flow of a mobile app. Our funnel showed that about 12 percent of new users never finished verifying their phone number, and the drop was much higher in regions with slower networks.

Task The team's goal that quarter was a new referral feature, but I believed these users were stuck at the very first step, so I asked for a week to understand why.

Action I tested the flow on a throttled network and on two older phones. The one-time code often arrived after our 30-second timer ran out, and the screen then told people to start over, which sent a second code and made the first one invalid. People were entering a code that had just stopped working. I worked backwards from what the user saw: I extended the timer to two minutes, kept the last two codes valid for a short window, added a "call me with the code" option after 45 seconds, and rewrote the error to say exactly what to do. I shipped it to a small share of users first and compared completion rates.

Result Verification completion rose from 88 to 96 percent, which meant thousands more new users each week actually got into the app, more than the referral feature was expected to bring in its first quarter. Support tickets about codes fell by more than half.

Tip: Show you looked at the problem from the user's screen, and that you chose their need over the planned roadmap with evidence. Customer focus: weak answer, follow-ups and level differences

OwnershipOwnership Tell me about a time you took on something outside your responsibilities.Tell me about a time you made a decision that hurt in the short term but helped in the long term. Write a story
Sample answer: Tell me about a time you took on something significant outside your area.

Act on behalf of the whole company and the long run, not just your team; never say "that is not my job".

Situation I launched a feature that let users export their order history. My ticket was closed and the launch looked fine on our team's dashboards.

Task A week later I noticed in a shared channel that the data warehouse team was seeing much higher load at night. I suspected my feature, even though nobody had connected the two.

Action I checked and found my export jobs ran at midnight in a single batch and read directly from tables the warehouse team also used for nightly loads. It was not on my roadmap any more, but it was my change. I told the warehouse team straight away that my launch was the likely cause, and agreed a temporary limit on export jobs while I fixed it properly. I then moved exports to read from a replica, spread jobs across the night, and added a dashboard both teams could see. I also wrote a short note for our launch checklist asking engineers to check shared database load before releasing batch features.

Result The warehouse team's nightly loads went back to finishing on time within a week, and exports got faster for users because the replica was less busy. The checklist item has since caught two similar issues before launch.

Tip: Pick a story where you owned the consequences of your work after it was "done", especially when they landed on another team. Ownership: weak answer, follow-ups and level differences

Invent and SimplifyInvention and simplification Tell me about a time you simplified a complex process or system.Describe an idea of yours that the team adopted. Write a story
Sample answer: Tell me about a time you found a simpler way to do something.

Look for new ideas everywhere, keep making things simpler, and be willing to be misunderstood for a while when trying something new.

Situation Our team managed about 400 feature flags in one hand-edited configuration file. Typos broke deployments a few times a month, and nobody knew which flags were still in use.

Task I wanted to stop the broken deployments and make the flag list trustworthy, without introducing a large new system.

Action Instead of buying a flag management product, I asked what we actually needed: a valid list, an owner and an expiry date for each flag. I wrote a small schema and a generator that turned it into typed code, so a misspelt flag became a compile error instead of a production problem. I added an owner and an "expires" field to every flag, and a weekly job that opened a ticket for expired flags. Some teammates worried this was extra process, so I migrated all existing flags myself and showed that adding a flag now took one line instead of three files.

Result Deployment failures from flag typos went to zero, and in two months we deleted 160 dead flags, which removed about 3,000 lines of unused code paths. Two other teams adopted the generator.

Tip: Show that you questioned what was really needed, and that the simple version removed work rather than adding a new system. Invention and simplification: weak answer, follow-ups and level differences

Are Right, A LotDiving deep · Ambiguity and judgment Tell me about a time you had to balance a customer need against a technical constraint.Describe a project where the requirements were unclear. How did you move forward? Write a story
Sample answer: Tell me about a time you made a decision with incomplete information that turned out to be right.

Have strong judgment and good instincts, seek out views that differ from yours, and work to prove your own beliefs wrong.

Situation My team needed a queue for a new background job system. Most of the team wanted to run our own open-source message broker because we had used it before.

Task I was asked to make the recommendation. I leaned towards a managed queue service, but I knew I could be biased, so I wanted a decision that would hold up.

Action I wrote down the three things that mattered most: operating effort, cost at our expected volume, and delivery guarantees. Then I asked the two strongest supporters of running our own broker to write the best case for it, and I took their points seriously. They were right that self-hosting was cheaper at very high volume. I checked our real numbers: we were about twenty times below the point where that mattered, and running our own broker meant on-call work for a team of four. I also listed what would make me change my mind, such as volume growing tenfold, and set a review date six months out.

Result We chose the managed queue. At the six-month review our volume was still far below the break-even point, and we had spent no on-call time on the queue. The written criteria made the review take one meeting.

Tip: Show how you looked for disagreement and data that could prove you wrong, not just that you guessed correctly. Ambiguity and judgment: weak answer, follow-ups and level differences

Learn and Be CuriousCuriosity and growth Tell me about the hardest feedback you have given or received.Tell me about a project that failed. What did you learn? Write a story
Sample answer: Tell me about something you learned on your own that improved your work.

Never stop learning, keep improving yourself, and be curious about new possibilities and how things work.

Situation A monthly finance report our team owned took six hours to run and sometimes timed out. The usual fix had been to give the database more memory.

Task I was curious why one query could be that slow, and I realised I did not really understand how the database chose to run queries.

Action Over two weeks I spent a few evenings learning how the query planner works: reading the documentation, then running the report's queries with the planner's explanation turned on. I found the planner was using a poor plan because our table statistics were badly out of date after a large data import, and one join was doing a full scan of 200 million rows. I refreshed the statistics, added one composite index and rewrote a subquery as a join. I wrote a one-page explainer with three examples from our own queries and ran a 30-minute session for the team.

Result The report went from six hours to twelve minutes and stopped timing out. Two teammates used the same method the next month to fix slow queries in their own services.

Tip: Show curiosity that led to a result at work, and ideally to others learning too. Curiosity and growth: weak answer, follow-ups and level differences

Hire and Develop the BestGrowing others Tell me about a time you helped a teammate succeed.Tell me about someone you mentored. What changed for them? Write a story
Sample answer: Tell me about a time you helped improve hiring or developed someone on your team.

Raise the level of talent with every hire and promotion, spot strong people, and take developing others seriously.

Situation Our team was hiring four engineers. Interviewers each asked their own favourite questions, and debriefs often ended in arguments because people had judged different things.

Task I was one of the regular interviewers, and I wanted our decisions to be fairer and better at predicting how people would do on the job.

Action I collected the questions from our last twenty interviews and grouped them by what they tested. We were asking a lot of puzzle-style questions and almost nothing about debugging or reading existing code, which was most of our real work. I proposed two changes: a debugging exercise using a small, realistic codebase, and a written rubric for each round that described what weak, good and strong looked like. I ran a practice session where five interviewers scored the same recorded mock interview, and we adjusted the rubric until our scores were close. I also offered to shadow and coach two newer interviewers.

Result Debriefs became shorter and more focused on evidence. Of the four engineers we hired that half-year, all were rated as meeting or exceeding expectations at their first review, and the two newer interviewers became fully independent within a month.

Tip: Show you raised the bar with a system, such as a rubric or a coaching plan, not just that you interviewed people. Growing others: weak answer, follow-ups and level differences

Insist on the Highest StandardsHigh standards Tell me about a time you were not satisfied with "good enough".Describe a time you had to choose between quality and speed. Write a story
Sample answer: Tell me about a time you refused to compromise on quality.

Keep raising the bar even when others think it is high enough, and do not let defects pass down the line.

Situation We were about to launch a feature that let users send money to friends inside our app. In the final review I noticed that if the app retried a request after a network timeout, the server could process the same transfer twice.

Task The launch date had been announced, and some on the team felt duplicates would be rare enough to fix later. I believed charging someone twice was not acceptable at any rate.

Action I estimated the risk with real data: our logs showed retries after timeouts on about 0.3 percent of requests, which would mean dozens of duplicate transfers a day. I proposed a small, specific fix rather than a delay of unknown length: every transfer request would carry a unique key from the app, and the server would store it and return the original result for any repeat. I wrote the change and its tests in two days, including a test that fired the same request twenty times in parallel. I asked for a three-day delay and explained exactly why.

Result We launched three days late with zero duplicate transfers in the first six months. The unique-key pattern became a rule for every endpoint that moves money, and we added a check for it in code review.

Tip: Quantify why the standard mattered and offer a concrete fix with a short, honest delay. High standards: weak answer, follow-ups and level differences

Think BigInvention and simplification · Big impact and long-term thinking Tell me about a time you made a decision that hurt in the short term but helped in the long term.Tell me about a time you simplified a complex process or system. Write a story
Sample answer: Tell me about a time you turned a small task into something much bigger.

Set a bold direction that inspires results, think differently, and look around corners for ways to serve customers.

Situation I built a small tool for my team that replayed a sample of real production requests against a test environment, so we could catch performance problems before release.

Task Other teams kept asking to borrow it. I saw a chance to turn a one-team script into a shared service that could protect every release in our organisation.

Action I wrote a short proposal describing the bigger version: any team could register their service, and every release would automatically be replayed against real traffic patterns with a clear pass or fail. I kept it grounded by listing the three incidents from the past year it would have caught. To prove it, I onboarded two more teams myself over a month while keeping my normal work, and measured what it found. With those results I asked for one more engineer for a quarter, and we built self-service sign-up, privacy filtering of request data and automatic reports.

Result Within two quarters, 22 services used it, and it caught 9 performance problems before they reached users. It became a standard step in our release process.

Tip: Show the bold idea, the proof you gathered before asking for more, and how many people it reached. Big impact and long-term thinking: weak answer, follow-ups and level differences

Bias for ActionSpeed and delivery Tell me about a time you missed a deadline.Tell me about a decision you made without all the data you wanted. Write a story
Sample answer: Tell me about a time you made a quick decision without all the information.

Move quickly, because speed matters and many decisions can be undone; take sensible risks rather than waiting for perfect information.

Situation I was on call when error rates on our search service started climbing on a Saturday evening. A deployment had gone out two hours earlier, but its changes looked unrelated to the errors.

Task I could spend time proving the deployment was the cause, or act. Errors were affecting about eight percent of searches and rising.

Action I asked myself whether a rollback could be undone easily, and it could; rolling back cost us nothing but a delayed feature. So I rolled back immediately and posted in the incident channel why, while I kept investigating. Errors dropped to normal within five minutes, which confirmed the link. I then found the cause: the deployment had updated a shared library that changed a default timeout. I wrote it up the next morning, and the feature was re-released on Monday with the timeout set explicitly.

Result Total impact was about 25 minutes instead of what could have been hours. I proposed and we adopted a simple rule: if errors rise within a few hours of a deploy, roll back first and investigate second.

Tip: Stress that the action was reversible, and that you explained it and followed up properly. Speed and delivery: weak answer, follow-ups and level differences

FrugalityDoing more with less Tell me about a time you achieved a lot with very few resources. Write a story
Sample answer: Tell me about a time you delivered something with limited resources.

Achieve more with less; limits push you to be resourceful and self-reliant, and spending more or adding people earns no extra credit.

Situation We needed a load-testing environment before a big sale event, but a full copy of production would have cost more than our team's monthly budget, and there was no time to request more.

Task I had to give the team a realistic load test within two weeks using what we already had.

Action I looked at what we really needed to test: the checkout path and its three dependencies, not the whole system. I built a scaled-down environment at one quarter of production size, using spare capacity in our staging account during nights only, and turned it off automatically each morning. To make results meaningful, I tested at one quarter of the target load and checked that resource use scaled roughly in a straight line, then confirmed the biggest risk, the database connection limit, with a targeted test at full load on a single component.

Result The tests cost about five percent of a full copy. They found a connection-pool limit that would have failed at about 60 percent of the expected peak. We fixed it, and checkout ran without errors through the sale.

Tip: Show how you cut the problem down to what mattered and still got a trustworthy result. Doing more with less: weak answer, follow-ups and level differences

Earn TrustIntegrity and trust Describe a mistake you made that reached production. How did you handle it?Tell me about a time you had to tell someone something they did not want to hear. Write a story
Sample answer: Tell me about a time you had to earn the trust of a team that doubted you.

Listen carefully, speak honestly, treat others with respect, and be willing to criticise yourself even when it is awkward.

Situation I moved to a platform team whose shared library was used by the mobile team. Two earlier releases had broken their app builds, and the mobile team had started copying our code instead of updating.

Task I was given ownership of the library. I needed the mobile team to trust our releases again, or they would keep falling behind on security fixes.

Action I asked their lead for a 30-minute meeting and mostly listened. They had three complaints: breaking changes without notice, release notes that hid what changed, and slow replies when things broke. I admitted we had got those wrong and did not argue the details. I made three small commitments I knew we could keep: a two-week notice for any breaking change, plain release notes written for app developers, and a reply within one working day. I added their build to our test pipeline so we would find breakages first. For two months I sent a short note after each release confirming their build had passed.

Result After three releases with no breakages, the mobile team moved back to our latest version and deleted their copied code. Their lead later asked us to review a design of theirs, which had not happened before.

Tip: Show listening, owning past mistakes without excuses, and small promises kept over time. Integrity and trust: weak answer, follow-ups and level differences

Dive DeepDiving deep · Ambiguity and judgment Tell me about a time you had to balance a customer need against a technical constraint.Describe a project where the requirements were unclear. How did you move forward? Write a story
Sample answer: Tell me about a time the data did not match what people were saying.

Stay close to the details, check them often, and be sceptical when numbers and stories do not match.

Situation After we redesigned our product page, our dashboard showed that "add to cart" clicks had risen by 15 percent. But sales had not changed at all, and the support team said users found the new page confusing.

Task The team wanted to call the redesign a success. I thought the numbers and the stories did not fit, and I wanted to know which was true.

Action I went to the raw click events instead of the dashboard. In the new page, the add-to-cart button sent an event when it was pressed, and the page also sent the same event when the cart opened. So one click often counted as two. I checked by replaying a sample of real sessions and counting actual clicks by hand. After correcting for the double count, add-to-cart was actually down three percent. I also watched recordings and found people were missing the size selector, which now sat below the fold on smaller phones.

Result We fixed the event, moved the size selector up and re-ran the test. Real add-to-cart then rose by four percent and sales followed. We added an automated check comparing event counts with server-side cart changes.

Tip: Show that a mismatch between numbers and anecdotes made you go to the raw data yourself. Diving deep: weak answer, follow-ups and level differences

Have Backbone; Disagree and CommitDisagreement and candor Tell me about a time you disagreed with your manager. What happened?Describe a decision you argued against but then had to carry out. Write a story
Sample answer: Tell me about a time you disagreed with a technical decision but committed to it anyway.

Respectfully challenge decisions you disagree with, even when it is uncomfortable, and once the decision is made, support it completely.

Situation A senior engineer on my team designed a new order-tracking feature in which the order service would call five other services synchronously for every request.

Task I believed this would make order tracking fail whenever any one of those services was slow. I needed to raise it clearly without derailing the project.

Action I wrote my concern in the design review document rather than in a hallway conversation, with numbers: if each service was available 99.9 percent of the time, the chain together would be about 99.5 percent, roughly three and a half hours of failures a month. I proposed an alternative where the order service kept a copy of the data, updated by events. The senior engineer explained their reasons: the event approach would take three more weeks, and the feature was a trial that might be removed. Our manager decided to go ahead with the synchronous design for the trial. I said I still had concerns, then committed. I added timeouts and fallbacks to each call so a slow service would show partial data instead of failing.

Result The trial launched on time and the fallbacks prevented two outages. When the feature became permanent, the team moved to the event-based design, and my original document was used as the starting point.

Tip: Show you disagreed openly with data, understood the other side, and then made the chosen plan work. Disagreement and candor: weak answer, follow-ups and level differences

Deliver ResultsSpeed and delivery Tell me about a time you missed a deadline.Tell me about a decision you made without all the data you wanted. Write a story
Sample answer: Tell me about a time you delivered an important result despite a setback.

Focus on the inputs that matter most and deliver them with the right quality and on time; when setbacks hit, rise to the challenge and do not settle.

Situation Our team had to move 40 million customer records from an old database to a new one before the old database's support contract ended, eight weeks away.

Task I led the migration. Three weeks in, our test run showed the copy would take eleven days, far longer than the weekend window we had planned.

Action Instead of asking for a later date, I changed the approach. I split the migration into a background copy that could run for days while the old database stayed live, and a short final step that only copied changes made since. I wrote a checker that compared record counts and samples between the databases each night, so we would know the copy was correct before switching. I shared a revised plan with my manager the same week, including the one risk I could not remove: a slower final step if traffic was unusually high. We ran two full rehearsals on staging.

Result We switched over in a 40-minute window, two weeks before the deadline, with no data errors found by the checker or by customers. The two-step method became our standard for large data moves.

Tip: Show the setback, how you changed the plan instead of the date, and the result delivered on time. Speed and delivery: weak answer, follow-ups and level differences

Strive to be Earth's Best EmployerBig impact and long-term thinking Tell me about a time you made a decision that hurt in the short term but helped in the long term.Describe an idea of yours that the team adopted. Write a story
Sample answer: Tell me about a time you made your team a better place to work.

Work every day to make the workplace safer, fairer, more inclusive and more rewarding, and help the people around you grow.

Situation Our on-call rotation was wearing people down. Engineers were paged at night about twice a week, often for alerts that needed no action, and two teammates had said in a retrospective that they dreaded their on-call weeks.

Task I was not the manager, but I volunteered to lead an effort to make on-call sustainable within a quarter.

Action I collected every page from the past three months and sorted them: about 60 percent needed no human action. I deleted or downgraded those alerts with the agreement of each service owner. For the rest, I wrote runbooks so that any engineer, including new joiners, could handle the most common pages. I also noticed that teammates in another time zone were always given the night shifts, so I proposed a rotation that followed working hours across both locations, with handovers written in a shared document. Finally, I set up a short monthly review where whoever had been on call could raise what was painful.

Result Night pages dropped from about eight a month to under two. In the next team survey, satisfaction with on-call rose from the lowest score to above average, and two newer engineers joined the rotation confidently in their second month.

Tip: Show a concrete change to people's day-to-day experience, including those easy to overlook. Collaboration and inclusion: weak answer, follow-ups and level differences

Success and Scale Bring Broad ResponsibilityBig impact and long-term thinking Tell me about a time you made a decision that hurt in the short term but helped in the long term.Describe an idea of yours that the team adopted. Write a story
Sample answer: Tell me about a time you considered the wider impact of something you built.

Remember that size brings influence; think about the wider effects of your decisions on users, communities and the world, and try to leave things better.

Situation I was building a feature that showed delivery drivers' live location to customers waiting for an order. The first design stored every location update permanently, because it might be useful for analytics later.

Task I was the engineer building the storage. I was uncomfortable keeping a detailed record of where every driver had been, long after it was needed.

Action I raised it with my product manager and our privacy team, framed around the drivers who had not been asked. I asked what the analytics team actually needed. They needed average delivery times by area, not exact routes. I proposed keeping exact locations only while an order was active, plus 24 hours for support questions, then storing only rounded, aggregated data. I also made sure customers stopped seeing a driver's location as soon as the delivery was complete. I wrote the deletion job, a test that proved old data was gone, and a short note explaining the decision for future engineers.

Result We launched with about 95 percent less stored location data, analytics still got everything it needed, and storage costs were lower. The retention rule was later reused for two other location features.

Tip: Show you thought about people affected beyond the immediate user, and turned that into a concrete design choice. Big impact and long-term thinking: weak answer, follow-ups and level differences

How to answer

Use STAR: the Situation in one or two sentences, the Task you owned, the Actions you took (in the first person, the longest part), and the Result with a number. One strong story can cover several principles, so write a few good ones in the story bank and tag them.

Behavioral questions by theme, with a tip for each · How the behavioral round is scored

Questions people ask

What is the Amazon behavioral interview based on?

Leadership Principles: Customer Obsession, Ownership, Invent and Simplify, Are Right, A Lot, Learn and Be Curious, Hire and Develop the Best, Insist on the Highest Standards, Think Big, Bias for Action, Frugality, Earn Trust, Dive Deep, Have Backbone; Disagree and Commit, Deliver Results, Strive to be Earth's Best Employer, Success and Scale Bring Broad Responsibility. Each interviewer gets two to four principles to test and asks STAR questions about them, even in coding and design rounds. Interviewers write down quotes as evidence and argue hire or no-hire in a debrief run by the Bar Raiser.

How do I answer Amazon questions on "Customer Obsession"?

Start from what the customer needs and work backwards; keep their trust even when a competitor or an internal goal pulls you elsewhere. Show you looked at the problem from the user's screen, and that you chose their need over the planned roadmap with evidence. A typical question: Tell me about a time you went out of your way for a customer.

How do I answer Amazon questions on "Ownership"?

Act on behalf of the whole company and the long run, not just your team; never say "that is not my job". Pick a story where you owned the consequences of your work after it was "done", especially when they landed on another team. A typical question: Tell me about a time you took on something significant outside your area.

How do I answer Amazon questions on "Invent and Simplify"?

Look for new ideas everywhere, keep making things simpler, and be willing to be misunderstood for a while when trying something new. Show that you questioned what was really needed, and that the simple version removed work rather than adding a new system. A typical question: Tell me about a time you found a simpler way to do something.