Amazon Interview Questions — Page 87

Amazon·Technical Program Manager·Onsite - Behavioral / Leadership

Jun 2026

Amazon TPM loop focused heavily on behavioral stuff around customer obsession and data-driven decision making. Three main question areas came up and the follow-ups were relentless, way more than I expected.

What methods have you used to gather customer feedback?
Tell me about a time you had to make a quick decision without enough data.
How do you deep-dive to solve a problem when something is going wrong?

“I went straight to surveys and NPS and the interviewer just kind of waited.”

View Post

Amazon·Software Engineer·Online Assessment (OA)

Jun 2026

Amazon SWE online assessment with a distribution center optimization problem. Pretty much a pure algorithmic question dressed up in supply chain language, which took me a minute to see through.

Given a sequence of daily demand values, find the minimum number of unique distribution hubs needed if hub selection follows the rule: pick a higher-numbered hub when demand increases, a lower-numbered hub when demand decreases, and reuse the same hub when demand stays flat.

“The Amazon branding around 'distribution centers' threw me off for longer than I'd like to admit.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

Jun 2026

Amazon ML engineer technical screen, went deep on normalization techniques in transformers. Pretty brutal if you haven't thought carefully about the math behind LayerNorm recently.

Explain Layer Normalization in transformers, including the equation, the roles of gamma and beta, and what epsilon is doing there.
Where is LayerNorm applied in a transformer block, and how does placing it before versus after the residual connection affect gradient flow and training stability?
How does LayerNorm differ from BatchNorm, and when would you prefer one over the other?
What is RMSNorm and how does it compare to LayerNorm in terms of computation and what it gives up?
What are the implications of LayerNorm placement and design choices on inference latency and memory usage?

“I wrote out the equation fine but fumbled explaining epsilon.”

View Post

Amazon·Software Engineer·Technical Phone Screen

Jun 2026

Interviewed for a security engineer role at Amazon, got asked a pretty foundational networking question that I probably over-explained.

What is the difference between HTTP and HTTPS?

“Talked about TLS, certificates, encryption in transit.”

View Post

Amazon·Software Engineer·Onsite - System Design / Architecture

Jun 2026

Amazon system design round for a software engineer role, one big open-ended question about building a GitHub-style developer community backend. The scope was wide enough that I kept second-guessing whether I was going too deep on one thing or not deep enough on another.

Design the backend for a GitHub-inspired developer community platform, covering user auth, social follow graph, posts with comments and likes, a home feed, and basic notifications.
How would you handle the data model and key indexes for this system, specifically for the feed and the follow graph?
Walk through your API design for the core features: auth, posting, feed retrieval, and notifications.
How would you scale this system for high read volume on the feed, and what failure modes concern you most?
What should be strongly consistent versus eventually consistent in this system, and why?

“The question sounds manageable until you realize how many sub-problems are hiding inside it.”

View Post

Amazon·Data Scientist·Onsite - Behavioral / Leadership

Jun 2026

Amazon Data Scientist behavioral loop, all four questions were leadership principle deep-dives. The whole thing felt like a structured storytelling exercise more than a technical screen, which I wasn't fully prepared for.

Tell me about a high-impact project you delivered under a hard deadline. What did you cut, what risks did you take, and what was the measurable outcome?
Describe a time you had to teach yourself a new skill under serious time pressure, say within six weeks. How did you structure your learning, and how did you prove you'd actually gotten good at it?
Walk me through a decision you made when you only had partial information, maybe around 70% confidence. How did you structure it as a reversible experiment, and what were your guardrails?
Tell me about a problem that was outside your official scope that you took full ownership of. What did you put in place to make sure it didn't happen again, and how did you measure that over time?

“This one I actually felt okay about.”

View Post

Amazon·Software Engineer·Onsite - Coding / Algorithms

Jun 2026

Amazon SWE coding round, one question the whole time: design an LFU cache from scratch. Not a warmup, not a follow-up. Just that.

Design an LFU (Least Frequently Used) cache that supports get and put operations in O(1) time, evicting the least frequently used entry when capacity is reached.

“I knew LRU cold but LFU tripped me up at first because you need to track frequency AND recency within the same frequency bucket.”

View Post

Amazon·Data Scientist·Technical Phone Screen

Jun 2026

Amazon data science interview for a real-estate pricing role. The whole thing was one big open-ended ML system design question that covered basically every stage of the modeling pipeline. More breadth than depth, but you still had to know your stuff at each step.

Walk through how you'd design a home price estimation model end-to-end, covering how you'd define success metrics, collect and split data, engineer features, choose and validate a model, deploy it, monitor it in production, and handle user complaints when predictions are wrong.

“This is a lot to cover in one question and I think I fumbled the ordering a bit.”

View Post

Amazon·Technical Program Manager·Onsite - System Design / Architecture

Jun 2026

TPM system design round at Amazon, single question about building a listening history pipeline. Pretty focused session, no behavioral stuff at all.

Design an application that tracks a user's song listening history, but only logs a play after the user has listened for more than 30 seconds. The data is intended for analytics purposes.

“The 30-second threshold is where I spent most of my time and I think that was the right call.”

View Post

Amazon·Software Engineer·Technical Phone Screen

Jun 2026

Amazon SWE coding screen, one problem about computing message latency by joining two CSV log files on a shared message ID. Pretty straightforward on the surface but the efficiency constraint made it more interesting than I expected.

Given two CSV log files, one recording when messages were sent and one recording when they were received, compute the latency for each message that appears in both files. Each row has a message type, a unique integer ID, and a timestamp. Output the message type, ID, and the difference in timestamps, sorted by ID. Aim for near-linear or O(n log n) runtime.

“My first instinct was a nested loop and I almost said it out loud before catching myself.”

View Post