Amazon Interview Questions — Page 56

Amazon·Software Engineer·Onsite - Behavioral / Leadership

May 2026

Amazon product designer interview, one question about career regrets. Not much else to go on but it's the kind of question that sounds easy until you're actually sitting there trying to answer it.

What do you regret in your design career?

“This one tripped me up more than it should have.”

View Post

Amazon·Software Engineer·Technical Phone Screen

May 2026

Amazon technical phone screen for a software engineer role, basically a deep knowledge check on hash maps. No LeetCode, just a conversation about internals. Felt more like a design discussion than a quiz, which I wasn't fully expecting.

How does a typical in-memory hash map work, and how would you handle collision resolution? What are the trade-offs between different strategies?
Walk me through load factor, when and how resizing and rehashing happens, and what the amortized complexity actually means.
How do you handle deletions in a hash map, particularly when using open addressing?
What does worst-case behavior look like for a hash map, and what are practical ways to mitigate it?
Does iteration order in a hash map have any guarantees, and what can cause it to change?
What are the concurrency concerns when multiple threads read from and write to a hash map, and what approaches exist for thread safety?

“I went with chaining vs open addressing and talked through cache locality.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

May 2026

Amazon ML engineer interview, got asked about data cleaning and prep. Pretty standard technical screen but the question had more depth to it than I expected going in.

How do you clean and prepare datasets that are messy or difficult to work with?

“I rambled a bit here.”

View Post

Amazon·Software Engineer·Onsite - Behavioral / Leadership

May 2026

Went through the Amazon SWE behavioral loop and basically had to live inside their leadership principles for the whole thing. Every question was a variation of 'prove you've done this before' with follow-ups on numbers and trade-offs. Not a bad experience but you really need your stories tight going in.

Tell me about a time you went above and beyond for a customer or end user.
Describe a situation where you took ownership of a problem that wasn't strictly your responsibility.
Give an example of a time you had to make a decision quickly without all the information you wanted.
Walk me through a time you dug deep into data or a technical problem to find a root cause others had missed.
Tell me about a time you simplified a complex process or system.
Describe a time you had to rebuild trust with a teammate or stakeholder after something went wrong.
Tell me about a time you raised the bar on quality in your team or project.
Give me an example of a meaningful failure and what you learned from it.

“I structured my answer as situation, action, result and it felt okay until they asked for the specific metric showing impact.”

View Post

Amazon·Machine Learning Engineer·Onsite - Coding / Algorithms

May 2026

Amazon MLE loop, first technical round. The whole session was a single four-part Morse code problem that escalated from basic encoding all the way to full backtracking enumeration. Pacing was the real challenge, not the algorithms themselves.

Given a string of letters, output its Morse code representation. Then, given a Morse string with delimiters between characters, recover the original text.
Given a vocabulary and a continuous Morse string with no delimiters, recover the single source word that produced it.
Extend the previous decoder to enumerate every possible sequence of vocabulary words whose concatenated Morse equals the input string, using backtracking.

“Parts 1 and 2 are basically table lookup plus a split call.”

View Post

Amazon·Data Scientist·Technical Phone Screen

May 2026

Amazon Data Scientist technical screen, heavy SQL and pandas. Two-part problem with a tricky de-duplication rule baked in that you had to carry through both parts consistently. Felt more like a take-home in spirit but delivered live.

Write a single SQL query (no temp tables, no correlated subqueries) that applies a same-day order de-duplication rule, then returns one row per kept order per customer with columns for order revenue, a 3-day rolling revenue window, and a dense rank by revenue over the last 7 days. Only customers with at least two completed orders in that window should appear.
Using pandas, produce a per-customer summary DataFrame for the last 7 days containing total revenue, the top product category by revenue (alphabetical tiebreak), and that category's share of total revenue. Apply the same same-day de-duplication rule before any aggregation, and do it without Python row-level loops.

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

View Post

Amazon·Software Engineer·Technical Phone Screen

May 2026

Amazon SWE interview with a greedy/simulation problem that looks deceptively approachable but has some real edge cases hiding in it. Nothing behavioral, just the one algorithmic question.

You have an array of positive integers representing product quantities. On day i (1-indexed), you must pick exactly i distinct products and take one unit from each. Quantities can't go below zero. What is the maximum number of days this process can continue?

“Took me a minute to even parse what was being asked.”

View Post

Amazon·Software Engineer·Technical Phone Screen

May 2026

Amazon SWE interview with a DP-heavy coding problem. Nothing too surprising but it required knowing your way around word break style recursion with memoization.

Given an array of strings, find all words in the array that can be formed by concatenating two or more other words from the same array.

“Classic word break variant but applied across the whole list, which trips you up if you just think about it as a single string problem.”

View Post

Amazon·Software Engineer·Technical Phone Screen

May 2026

Amazon SWE coding round, one question the whole time, grid-based flow problem that's basically a twist on a classic LC problem. Not the hardest thing I've done but I fumbled the setup more than I'd like to admit.

Given an m x n grid where each cell has a height value, water flows from a cell to any adjacent cell (4-directional) with height less than or equal to the current cell. Water is dropped on a specific set of starting points and flows until it exits the grid. Return all cells that end up wet.

“I recognized it was related to the Pacific Atlantic problem pretty fast, which helped.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

May 2026

Deep learning fundamentals interview at Amazon for an ML Engineer role. Every topic had a follow-up that went two or three levels deeper than I expected, so if you're prepping, don't stop at the conceptual layer.

Explain FlashAttention: what problem it solves compared to standard attention, how the IO-aware tiling works, and what the memory and speed tradeoffs look like in practice.
Walk through LoRA: the low-rank decomposition idea, where adapters plug into a transformer, how rank and alpha interact, and how you'd actually serve or merge adapters at inference time.
Describe the RNN forward and backward pass, explain vanishing and exploding gradients, and compare LSTM and GRU gating. When would you still reach for an RNN over a transformer?
Explain the RLHF pipeline: how the reward model gets trained, how PPO is applied to fine-tune the policy, what the KL constraint against the SFT policy is doing, and what tends to go wrong.
Compare GRPO and GSPO directly: what each method averages over, why GSPO uses sequence-level importance ratios instead of token-level ones, and what that means for training stability.

“I started with the quadratic memory issue in vanilla attention and thought that was enough.”

View Post