The short answer
An Amazon-style mock interview is a practice round that pairs a technical problem, usually coding or system design, with behavioural questions built around Amazon's published Leadership Principles, the way Amazon's own interview guidance describes. On Revvox each round is 60 minutes live with an experienced engineer, scored 1 to 5 on five skills, with a written report within 24 hours. We are not affiliated with Amazon and do not claim our interviewers work there.
What an Amazon-style loop usually includes
Amazon publishes its Leadership Principles on its jobs site and tells candidates to expect questions about them. For software development engineers, the loop typically starts with an online assessment or phone screen, followed by a set of interviews. Most of those rounds mix a technical problem with one or two behavioural questions, each tied to a principle such as Customer Obsession, Ownership, Dive Deep or Deliver Results.
One interviewer in the loop is usually a Bar Raiser, someone from outside the hiring team whose job, as it is widely described, is to keep the hiring standard high across the company. Their round often leans heavily on behavioural questions and follow-ups.
Amazon's guidance asks for answers in the STAR shape: situation, task, action, result. Interviewers dig into the details, so a story only works if it is true, specific and yours. Formats differ by level, team and country; your recruiter has the details of your loop.
How it is scored
After the 60-minute round, the interviewer marks each line below 1 to 5 in your written report, due within 24 hours, with notes on what you said and wrote behind each mark. The verdict is one of Strong hire, Hire, Lean hire, Lean no hire or No hire.
| Skill (1 to 5) | What the interviewer looks for |
|---|---|
| Problem solving | A working approach reached methodically, with a better one when the interviewer asks for it. |
| Code quality | Clean, complete code with sensible names; no pseudo-code left where real code was asked for. |
| Testing and edge cases | You walk through edge cases unprompted and fix what you find. |
| Complexity | Correct time and space cost, and a reason to prefer one solution over another. |
| Communication | You explain as you go and keep the round on time, leaving room for the principle questions. |
| Skill (1 to 5) | What the interviewer looks for |
|---|---|
| Structure | STAR answers: a clear situation and task, then your actions, then a result, in about three minutes. |
| Ownership | You took responsibility beyond your own task and say what you decided, not what you were told. |
| Collaboration | You disagreed and committed, or earned trust, with a real example and its outcome. |
| Impact | A result with numbers or a clear before and after, tied to customers where it can be. |
| Self-awareness | You answer follow-ups honestly, including what went wrong and what you learnt. |
| Skill (1 to 5) | What the interviewer looks for |
|---|---|
| Requirements and estimates | You start from the customer and turn the use case into numbers before designing. |
| High-level design | A simple design that works, with a reason for each service and queue. |
| Data and storage | Storage chosen for the access pattern, with keys and growth thought through. |
| Scale and trade-offs | You plan for failure and peaks, and say what you would simplify to ship. |
| Communication | You drive the design and take follow-ups calmly. |
Practice questions in an Amazon style
Written by us in the style of these rounds, not taken from real interviews. Behavioural questions come with follow-ups; prepare for at least two each.
- Given orders arriving with timestamps, return the top k products by sales in any rolling hour.A heap with a sliding window, and how you expire old orders.
- Find the shortest delivery route between two warehouses on a grid with blocked cells.Breadth-first search, a clear visited set, and what changes if moves have different costs.
- Design a system that tells customers when an out-of-stock item is back.Subscriptions, a queue for stock events and avoiding a flood of messages when stock returns.
- Model a locker system where parcels of different sizes go into free lockers.Objects and their responsibilities, and how you would add a new locker size.
- Tell me about a time you went beyond your role to fix something for a customer.Customer focus and ownership, with what you did and what changed.
- Describe a decision you made without all the data you wanted.Bias for action, how you judged the risk and what you checked afterwards.
- Tell me about a time you disagreed with your manager and still committed.Raising the concern clearly, then committing and delivering.
How to rehearse it with a Revvox package
Most people take the Standard package: three 60-minute rounds. Because every Amazon-style round carries behavioural questions, ask for them in each round, not only the last. You get a report after every round, including notes on which stories held up under follow-ups and which did not.
- Round one: coding plus principles. A data structures problem, then two questions on principles such as Ownership or Dive Deep.
- Round two: design plus principles. System design for SDE II and above, or object-oriented design for SDE I, with principle questions at the end.
- Round three: a Bar Raiser-style round. Mostly behavioural, with deep follow-ups on two or three stories, plus a short technical question.
How to prepare
- Read the Leadership Principles on Amazon's jobs site and write one or two true stories for each of the ones that come up most for engineers.
- Write each story in STAR form, with numbers you can stand behind and what you personally did, not what the team did.
- Prepare for follow-ups: what would you do differently, what data did you look at, who disagreed with you.
- Keep coding sharp on arrays, hash maps, trees, graphs and heaps, and explain the cost of each solution.
- For SDE II and above, practise one system design end to end in 40 minutes.
Common mistakes in Amazon-style rounds
- Saying "we" all the way through. The interviewer needs to know what you did.
- Stories with no result, or a result you cannot measure in any way.
- Reusing one story for every principle. Interviewers compare notes.
- Strong stories but weak code, or the other way round. Both are scored in the same round.