← Back to Directory

Amazon interview notes, page 46 of 159

FAANG / Big Tech

Most of these Amazon write-ups come from Software Engineers (1,092), then Data Scientists (182) and Product Managers (165). The most reported rounds are the technical phone screen (618), the behavioral / leadership round (309), the online assessment (228) and the system design round (159).

1581 interview notes · updated Oct 2026

Amazon·Data Scientist·Onsite - Behavioral / Leadership

May 2026
Amazon Data Scientist behavioral loop. The whole thing is basically a leadership principles gauntlet dressed up as an interview, and they want numbers attached to every story you tell.
  • Tell me about a time you went above and beyond for a customer. What metric moved, and by how much?
  • Describe a situation where you took ownership of something outside your defined scope. What was the cost or trade-off you accepted?
  • Walk me through a time you missed a deadline. What would you do differently to prevent it from happening again?
  • Tell me about a time you received difficult negative feedback. How did you respond?
  • Describe a time you disagreed with a decision but ultimately committed to it. How did you escalate or de-risk before committing?
  • How have you handled ambiguity or conflicting goals between stakeholders, such as a PM pushing in one direction while data pointed another way?

“The quantification part is where most people stumble, including me.”

View Post

Amazon·Product Manager·Onsite - Behavioral / Leadership

May 2026
Amazon PM interview, behavioral round focused on leadership principles. Just the one question but it took the whole session to unpack.
  • Describe a situation where you proposed a solution that seemed counterintuitive, and walk through how you recognized that conventional thinking wasn't going to work.

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

View Post

Amazon·Software Engineer·Onsite - System Design / Architecture

May 2026
Amazon system design round focused entirely on transformer architecture internals. Three questions, all deeply technical, no behavioral fluff. Felt more like a machine learning research interview than a software engineering one, which I was not expecting.
  • Walk through scaled dot-product self-attention from scratch: define the query, key, and value matrices, explain the scaling factor and why it exists, describe how masking works, and walk through the softmax step. Then discuss the computational complexity.
  • Explain the full Transformer encoder and decoder architecture, including how residual connections are used, where normalization is applied, and how positional information is injected into the model.
  • Compare BERT and GPT in terms of pretraining objectives, architectural differences, and how each is typically used for downstream tasks. How do those design choices affect transfer learning and inference behavior?

“I knew this cold but fumbled the complexity part.”

View Post

Amazon·Software Engineer·Onsite - Coding / Algorithms

May 2026
Amazon SWE coding round with a string manipulation problem that looks deceptively simple until you actually try to figure out the minimum flips needed.
  • You're given a binary string (containing only 0s and 1s). A 'good' string is one that has no subsequence matching the pattern '010' or '101'. Each operation lets you flip a single character (0 to 1 or 1 to 0). What's the minimum number of operations to make the string 'good'?

“My first instinct was to scan for those patterns and greedily flip, but that falls apart pretty fast because fixing one occurrence can create another.”

View Post

Amazon·Software Engineer·Online Assessment (OA)

May 2026
Took the Amazon SDE 1 online assessment as a 2026 grad. Two coding questions: nailed the first, but the AI-assisted bug fix section was rough with only 1 out of 6 test cases passing. Waiting to hear back and not feeling great about it.
  • Solve a coding problem where all test cases must pass (standard algorithmic problem, first of two in the OA).
  • Given a piece of buggy code, identify and fix the bugs with AI assistance (second coding problem in the OA).

“Went fine, 15/15.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

May 2026
Amazon ML Engineer technical screen, and they went straight for a classic graph problem. Nothing too wild but the follow-up on complexity tripped me up a bit.
  • Given two strings and a word list, find the minimum number of single-character transformations to get from the first string to the second, where every intermediate string must exist in the word list. Return -1 if it's not possible.

“Classic BFS shortest path setup.”

View Post

Amazon·Data Scientist·Technical Phone Screen

May 2026
Amazon data scientist interview with a SQL-heavy question around customer segmentation and ranking. Pretty standard stuff but the tie-breaking follow-up is where things get interesting.
  • Given a transactions table, filter customers based on campaign criteria and then select exactly the top 5 by total order count. How do you handle ties when more than 5 customers qualify for that top-5 cutoff?

“The filtering and aggregation part was fine, GROUP BY user_id, ORDER BY order_cnt DESC, LIMIT 5, done.”

View Post

Amazon·Software Engineer·Technical Phone Screen

May 2026
Amazon SWE coding round, one question, sorting problem with a twist on the comparator. Not too bad but the custom sort logic took me a minute to get right.
  • Given an integer array where each element is an error code, sort the array so that less frequently occurring codes appear first. If two codes have the same frequency, the smaller value comes first. Return the full sorted array including duplicates.

“I jumped straight to sorting without thinking through the comparator and had to backtrack.”

View Post

Amazon·Software Engineer·Onsite - Behavioral / Leadership

May 2026
Amazon behavioral screen, pretty much just one question about customer pain points. Standard leadership principles stuff, nothing unexpected.
  • Can you walk me through a time you identified and resolved a significant pain point for customers?

“Classic Amazon question, ties straight into their customer obsession principle.”

View Post

Amazon·Software Engineer·Online Assessment (OA)

May 2026
Amazon SWE online assessment, second question was a fullstack debugging task on HackerRank. The setup was a movie watchlist app with 6 failing unit tests to fix under time pressure. Small screen, weak AI assistant, and not a great time.
  • You're given a partially broken fullstack project (a movie watchlist web app). Six unit tests are failing. Debug and fix the code so the tests pass.

“The failing tests covered stuff like adding a movie to a list, handling duplicates, deleting movies, bulk add then delete one by one, and then the edge cases where the list doesn't even exist.”

View Post