Amazon Interview Questions — Page 112

Amazon·Software Engineer·Onsite - Behavioral / Leadership

Jun 2026

Amazon program manager interview, behavioral round focused on data-driven decision making. Pretty standard LP territory but the follow-up questions had some bite to them.

Tell me about a time when data helped you make a decision. How did you gather and interpret that data, what trade-offs did you weigh, and how did you make sure your own assumptions weren't quietly steering the outcome?

“The base question felt manageable but the bias follow-up tripped me up a bit.”

View Post

Amazon·Software Engineer·Technical Phone Screen

Jun 2026

Amazon SWE coding round with a custom OOD/algorithm hybrid that caught me more off guard than I expected. The problem sounded simple at first but the design layer on top of the heap mechanics is where things get interesting.

Design a streaming music player class that accepts batches of songs per user, always plays the highest-frequency unplayed song next, and resets the played history once every distinct song has been played. Discuss the complexity of both the ingest and next operations.
How would you handle tie-breaking when multiple songs share the same frequency in the next() call?
Walk through the tradeoffs between using a frequency hashmap with a max-heap versus a bucket-array approach similar to an LFU cache for this problem.

“I went straight to the heap and kind of glossed over the class structure, which I think was a mistake.”

View Post

Amazon·Software Engineer·Online Assessment (OA)

Jun 2026

Amazon SWE coding round with a graph/greedy problem on a circular hub network. The problem statement was a bit vague on the edge weight definition since the screenshot cut off the constraints and expected output, which made it harder to fully verify a solution.

A drone starts at hub 1 on a circular network of m hubs and must visit a sequence of requested hubs in order. Given travel times between adjacent hubs, compute the minimum total travel time to fulfill all requests.

“The circular structure is the key part.”

View Post

Amazon·Data Scientist·Technical Phone Screen

Jun 2026

Amazon Data Scientist technical screen, heavy on Python and NLP implementation. The questions were layered and kept building on each other, which I wasn't fully prepared for.

Given a 50GB UTF-8 text file on disk, how would you stream it in Python without loading it into memory, count case-insensitive word frequencies with Unicode normalization, and emit the top 10 words with counts and percentage of total tokens?
What are the time and memory complexity of your streaming word count solution, and what practical optimizations would you apply?
Re-implement the word frequency counter using only plain Python dicts, without using collections.Counter.
Implement safe_min and safe_max functions that handle NaN values, mixed comparable types via a key function, return a default when the iterable is empty, raise ValueError when no default is given, and are stable with equal keys.

“This is where I spent most of my mental energy.”

View Post

Amazon·Data Scientist·Technical Phone Screen

Jun 2026

Amazon DS interview that was basically a two-hour SQL and causal inference gauntlet. The main question was a massive difference-in-differences setup with household spillover exclusions, propensity score matching, and a Python component all baked into one prompt. Not a casual screen.

Given a schema with users, reminders, CSAT scores, and session data, write SQL to build a user-week panel for a difference-in-differences analysis around first reminder exposure. Requirements include defining each user's first reminder date as their treatment date, keeping only an 8-week window around that date, excluding users whose household has an earlier-treated member, and outputting weekly aggregates of CSAT, session minutes, and purchases alongside treatment indicators.
Using the user-week panel you built, write a query that computes group-level pre and post means for treated and control users, then derives the 2x2 difference-in-differences estimate.
In Python, outline code to estimate propensity scores using logistic regression on baseline covariates like device, country, and pre-period usage, then perform 1:1 nearest-neighbor matching with a caliper of 0.05 and report standardized mean differences before and after matching.

“This was the bulk of the interview.”

View Post

Amazon·Software Engineer·Onsite - Coding / Algorithms

Jun 2026

Interviewed for a senior SDE role at Amazon AWS and came out the other side with a downgrade offer. The DSA round went sideways and they came back saying they'd only go junior level, which now has me wondering whether taking a step down at a big name is worth it given everything going on with layoffs and PIPs over there.

Solve a coding problem involving data structures and algorithms under interview conditions.

“This is where it fell apart.”

View Post

Amazon·Product Manager·Onsite - Behavioral / Leadership

Jun 2026

Amazon PM interview, behavioral round focused on customer obsession. Just the one question but they really wanted to dig into the details of what you actually did versus what your team did.

Describe a time you took customer feedback and used it to drive a meaningful product change or innovation.

“I had a story ready but fumbled the 'innovation' part.”

View Post

Amazon·Product Manager·Onsite - System Design / Architecture

Jun 2026

Product Manager system design round at Amazon, centered entirely on a hotel keycard migration case. It was one of those questions that sounds operational but quickly spirals into crypto, failure modes, and stakeholder tradeoffs all at once.

A hotel is moving from metal keys to electronic keycards. What are the strategic and customer-driven reasons for making this switch, and what benefits does it create for guests, staff, and the business?
What are the must-have features for an end-to-end keycard solution, covering the check-in experience, lost card handling, access controls, auditing, and integration with loyalty or mobile apps?
Describe a high-level system architecture connecting keycards, door locks, on-premise or cloud servers, and hotel management software. How do the card and lock validate each other?
What security measures would you put in place to defend against both physical attacks like card cloning and network-based attacks like packet interception or brute-force entry?
How would your system handle failure scenarios like a power outage, server downtime, or a card being used repeatedly in quick succession?

“I led with the guest angle (faster check-in, no front desk queue) and worked toward operational stuff like room turnover speed and front desk labor costs.”

View Post

Amazon·Data Scientist·Technical Phone Screen

Jun 2026

Amazon data science technical screen, heavy on ML fundamentals. The questions were pretty textbook but there were enough moving parts that it was easy to lose track mid-answer. Left feeling like I half-nailed it.

Define precision and recall, and explain when you would prioritize one over the other.
Compare L0, L1, L2, and L-infinity regularization. What are their differences and when would you use each?
What are the core assumptions underlying linear regression?
Write out the formulas for logistic regression and linear regression, and explain how the two models differ conceptually.

“Started okay, defined both fine.”

View Post

Amazon·Software Engineer·Technical Phone Screen

Jun 2026

Amazon SWE interview with a graph validation problem that looked straightforward until the follow-ups hit. Solid problem overall, felt like a fair test of whether you actually understand what you're doing vs just pattern-matching to a known algorithm.

Write a validation function for a course catalog where each course has a list of prerequisites. The catalog is valid only if there are no circular dependencies and every referenced prerequisite actually exists in the catalog. Return whether the catalog is valid, and optionally identify the offending course or cycle.
What is the time and space complexity of your validation solution?
What are the differences between using BFS and DFS for this catalog validation, and when would you prefer one over the other?

“I jumped straight to cycle detection and forgot about the second condition entirely, that every referenced prerequisite needs to exist in the catalog.”

View Post