← Back to Directory

Amazon interview notes, page 139 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·Machine Learning Engineer·Technical Phone Screen

Jul 2026
Amazon ML Engineer screen, one meaty technical question about regularization that sounds deceptively textbook until you're actually in it trying to explain the tradeoffs out loud.
  • What is overfitting, why is it a problem, and how do L1 vs L2 regularization address it? Walk through how each modifies the loss function, what effect each has on the model weights, and when you'd choose one over the other.

“I knew this cold on paper but fumbled the 'when would you prefer one over the other' part.”

View Post

Amazon·Software Engineer·Technical Phone Screen

Jul 2026
Amazon SWE interview that went deep into probabilistic algorithms. One question, but it had a lot of moving parts and I felt underprepared for the complexity analysis portion.
  • You're building a lottery system where each participant's chance of winning scales with how much they've spent. Given a list of (participant_id, spending) pairs and an integer k, how would you select k distinct winners with probability proportional to spending? Walk through your algorithm, why it's correct, and its complexity.

“I went straight for the prefix-sum plus binary search approach.”

View Post

Amazon·Software Engineer·Onsite - Product Sense / Strategy

Jul 2026
Amazon SWE interview with a product strategy scenario about build vs buy. One question, but it had a few layers to it that I didn't fully appreciate until after.
  • You're evaluating a third-party tracking platform against building one in-house. How do you weigh the tradeoffs, and which factors matter most?

“I went straight to cost and kind of ignored the launch timeline angle, which in hindsight was probably the most important piece.”

View Post

Amazon·Data Scientist·Technical Phone Screen

Jul 2026
Amazon DS interview focused on a building-operations case about elevator allocation. The whole thing was one big experiment design question with follow-ups that kept adding wrinkles. Pretty involved for a single question session.
  • A building owner wants to reserve one out of six elevators exclusively for food-delivery couriers to reduce resident wait times. Design an A/B test to evaluate this policy and define how you would measure success.
  • If only 8 towers are available and meal-time peaks dominate the signal, would you still run a between-building test or switch to a switchback design? Walk through the trade-offs.
  • Compliance is only around 60% because many couriers ignore the reserved elevator. How does this affect your estimate, and how would you distinguish between intent-to-treat and treatment-on-the-treated effects?
  • Suppose resident wait time improves but courier delivery times get significantly worse, leading to complaints on the delivery app that damage the building's reputation. How do you frame that trade-off and make a recommendation?

“The first thing I fumbled was jumping straight to metric definition without naming who the primary customer even was.”

View Post

Amazon·Software Engineer·Online Assessment (OA)

Jul 2026
Amazon SWE coding round with a queue simulation problem that had a follow-up twist. The core task was straightforward enough but the ring variant caught me off guard and I spent way too long second-guessing my recursion.
  • Given N students in a queue, where each student at the front can either print and leave or move to the back, enumerate all valid completion orders (permutations of length N).
  • Follow-up: the queue becomes a circular ring with a moving pointer. At each step you can either print and remove the current student (advancing the pointer to the next remaining one) or skip and advance. Enumerate all valid removal orders.

“Classic backtracking setup once you see it, but my first instinct was to think about it iteratively and I wasted a few minutes going down that path.”

View Post

Amazon·Software Engineer·Technical Phone Screen

Jul 2026
Amazon SWE coding round, one problem the whole time. Seemed straightforward at first but the commit semantics tripped me up more than I expected.
  • Design and implement a Layer class that supports put(key, value), get(key), commit_batch(), and get_state() operations, where commit_batch collapses all current state into a permanent snapshot and discards prior intermediate states. Optimize for both performance and memory given up to 10,000 operations.

“The put and get parts were fine, like five minutes tops.”

View Post

Amazon·Software Engineer·Online Assessment (OA)

Jul 2026
Took the Amazon new grad SDE-1 OA and it was a mixed bag. First coding question went pretty well but the AI portion was a disaster, kept looping on the same output for like half an hour.
  • Coding problem 1 on the online assessment.
  • AI-related coding or analysis problem on the online assessment.

“Got 14 out of 15 test cases which felt decent.”

View Post

Amazon·Data Scientist·Onsite - Cross-functional / Panel

Jul 2026
Amazon Data Scientist interview with a meaty cross-functional scenario question about choosing between vendor options for a training program. One question, but it had five sub-parts that basically covered everything from stakeholder wrangling to writing a decision memo to contingency planning. More like a mini case study than a standard interview question.
  • You need to drive a build-vs-buy decision across three options (two external vendors and an internal build) for a training program where vendor solutions store employee data externally. Walk through how you'd align stakeholders across HR, Security, Legal, and Finance; set hard guardrails vs. negotiables; communicate trade-offs and produce a decision memo; plan contingencies for timeline slippage or a mid-project security issue; and justify your final recommendation with 90-day success metrics.

“This one took me a second to even parse.”

View Post

Amazon·Software Engineer·Online Assessment (OA)

Jul 2026
Amazon SWE coding round with a dynamic programming problem on partitioning arrays. The problem was dressed up in server/expense language but once you strip that away it's a pretty classic min/max partition cost question.
  • Given a list of servers with associated costs and an integer p, partition the list into exactly p contiguous segments. The cost of each segment is the sum of its first and last element. Return both the minimum and maximum total cost achievable across all valid partitions.

“Took me a minute to realize the segment cost formula means the endpoints are what matter, not the whole interior.”

View Post

Amazon·Product Manager·Onsite - Product Sense / Strategy

Jul 2026
Amazon PM interview with a system design and product strategy question about building a recommendation engine. Pretty open-ended, which I wasn't fully prepared for.
  • How would you design a recommendation system for Amazon, and what metrics would you use to measure whether it's working?

“I jumped straight into the algorithm side of things and the interviewer kind of steered me back toward the product framing.”

View Post