Amazon Interview Questions — Page 28

Amazon·Software Engineer·Online Assessment (OA)

Apr 2026

Amazon SWE online assessment with at least two algorithm problems in the same window. The main one was a ring-based redistribution problem with a directional constraint that trips you up if you approach it naively. Prep on the line version first or you'll probably get burned.

You have n warehouses arranged in a circle, each holding some number of products. Moving one product across a single edge costs 1, but all moves must flow in one fixed direction around the ring (no mixing clockwise and counterclockwise). Given that the total is divisible by n, find the minimum cost to equalize all positions.

“The directional constraint is what makes this hard.”

View Post

Amazon·Software Engineer·Onsite - Behavioral / Leadership

Apr 2026

Behavioral round for a software engineering internship at Amazon, about 25-30 minutes with follow-ups that can go pretty much anywhere depending on what you say. Nothing too surprising in terms of question topics but the follow-ups are where it gets real.

Why do you want to work at Amazon specifically?
Why are you interested in this internship program or this particular role?
Tell me about a time you had to make a quick decision under pressure.
Describe a situation where you dealt with a deadline, whether it changed, was at risk, or required you to make tough prioritization calls.

“I had an answer prepped but it felt a little rehearsed coming out.”

View Post

Amazon·Software Engineer·Technical Phone Screen

Apr 2026

Amazon SWE interview with a frontend component design question. Pretty straightforward on the surface but there's more to it than just slapping some divs together.

Design and implement a reusable dialog component that has a background color, title, close button, and action buttons.

“My first instinct was to just build a functional modal and call it done.”

View Post

Amazon·Data Scientist·Onsite - System Design / Architecture

Apr 2026

Amazon data scientist loop with a heavy focus on data modeling and SQL. The whole session felt like a database design exam more than anything else. Walked out unsure if I'd been thorough enough on the warehouse side.

Design a normalized OLTP schema for a university domain covering Students, Courses, Departments, Instructors, and Enrollments. Include primary and foreign keys, constraints like preventing a student from enrolling in two sections of the same course in the same term, a controlled grade value set, and the indexes you would add and why.
Design a star schema data warehouse for enrollment analytics. Define the grain of FactEnrollment, the measures it should carry, the dimension tables, surrogate key strategy, and how you would implement SCD Type 2 on the student dimension to track major changes with point-in-time correctness.
How would you partition and cluster a FactEnrollment table that has over 100 million rows? Justify your choices based on typical query patterns and maintenance considerations.
Write a SQL query to find all students who enrolled in both 'CS101' and 'MATH201' during the same term.
For the most recently completed term, write a SQL query returning the top 3 departments ranked by credit-weighted average GPA. Ties should be broken deterministically.
Compare the trade-offs between OLTP and OLAP database design in this university scenario, and explain when you would choose a snowflake schema over a star schema.

“I went straight to the Enrollment table and started listing columns before thinking about the constraint properly.”

View Post

Amazon·Data Scientist·Onsite - Behavioral / Leadership

Apr 2026

Amazon Data Scientist loop, heavy on leadership principles with a clear expectation that every answer comes with numbers and a before/after story. The questions are layered and they will push you on the 'so what' if your impact isn't quantified. Prepare more than you think you need to.

Tell me about a time you redefined or completely overhauled a metric that ended up changing your team's priorities. How did you spot the flaw in the original metric, validate the replacement statistically, and get the team to actually stick with it?
Describe a situation where you strongly disagreed with your manager and peers but went ahead with their decision anyway. What data did you bring, what risks did you flag, and what did the outcome look like?
Walk me through a case where you had to dig through multiple layers, data, pipeline, business process, to find a root cause that wasn't obvious at first. How did you rule out other explanations and measure the impact of the fix?
Give an example where you took an unpopular position and it actually changed the outcome of a decision. How did you stay firm without alienating people, and how did you rebuild trust afterward?
Tell me about a project that hit serious unexpected obstacles, whether resourcing, data quality, or shifting goals. How did you re-scope, manage the risk, and still deliver something meaningful?
When your direct customer is an internal stakeholder, how did you connect the value of your model all the way to the end user? What was the chain from your output to actual customer benefit, and how did you verify it?
Describe a time you shipped something under a hard deadline without having all the information you wanted. What guardrails or rollback plans did you put in place?
Tell me about a time you took on significant work outside your normal responsibilities to unblock your team. How did you balance that against your own goals without burning out?
Tell me about a piece of tough feedback you received. How did you figure out if it was valid, what did you change, and how did you show that the change actually stuck?
What's the most interesting data or ML project you led that actually moved a business metric? And if budget were no constraint, how would you scale its impact by 10x, including team, data, tooling, and what risks you'd need to manage?

“This one is deceptively hard.”

View Post

Amazon·Product Manager·Onsite - Behavioral / Leadership

Apr 2026

Interviewed for a PM role at Amazon, and the whole thing basically came down to how you handle not knowing what you're supposed to do.

How do you handle situations where the problem or path forward isn't clearly defined?

“This is basically Amazon's way of asking if you freeze up or take action when things are messy.”

View Post

Amazon·Product Manager·Onsite - Behavioral / Leadership

Apr 2026

Amazon PM loop, behavioral focus. Just the one question shared but it's a heavy one, the kind where you really need a story that shows actual organizational impact, not just team-level stuff.

Describe a situation where your actions shifted the direction of the broader organization.

“This one tripped me up more than I expected.”

View Post

Amazon·Software Engineer·Onsite - Behavioral / Leadership

Apr 2026

Behavioral round at Amazon for a SWE role, basically one question but it carried the whole conversation.

Tell me about a time when you significantly exceeded expectations on a project or task.

“I had a story prepped but it felt thin the moment I started telling it.”

View Post

Amazon·Software Engineer·Hiring Manager Screen

Apr 2026

Amazon SWE interview with a scenario-based question about deployment quality and testing practices. Pretty situational, felt more like a judgment call exercise than a technical deep dive.

Your manager brings up a recurring problem where code gets deployed without sufficient testing. She asks you to rank several possible solutions by effectiveness. How do you order them and why?

“The scenario gave four options to rank: automated gates that block untested code from deploying, written SOPs for testing requirements, a root-cause analysis report, and troubleshooting guides for code that's already out in the wild.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

Apr 2026

Amazon ML engineer interview with a dynamic programming batch optimization problem. The question was more involved than I expected for a phone screen, required both an optimal answer and reconstructing the actual partition.

Given an array of K document lengths and G available GPUs, partition the documents into at most G non-empty batches to minimize total padding waste, where padding waste is defined as the sum over batches of (batch size times max length in batch minus sum of lengths in batch). Return the minimum waste and one valid partition. Handle edge cases where K is 0 or G is at least K.

“I knew sorting was the right first move since grouping similar lengths reduces padding, but formalizing the DP recurrence took me longer than I'd like to admit.”

View Post