Amazon Machine Learning Engineer interview questions

101 Machine Learning Engineers have shared notes from their Amazon interviews, with the questions they were asked in each round. That is 6% of the 1,581 Amazon interview notes on Screna AI.

101 Machine Learning Engineer interview notes

Amazon·Machine Learning Engineer·Technical Phone Screen

Jul 2026
Amazon ML Engineer interview with a tree-traversal problem dressed up as an org chart question. Clean problem statement but the tie-breaking rule is the kind of thing that'll burn you if you're not paying attention.
  • Given a company org chart represented as a list of employee records (each with an id, name, salary, and manager_id), find the manager whose team of all direct and indirect reports has the highest average salary. Break ties by returning the smaller id. Return -1 if no managers exist.

“The core traversal isn't the hard part.”

View Post

Amazon·Machine Learning Engineer·Onsite - Behavioral / Leadership

Jul 2026
Behavioral loop for an ML Engineer role at Amazon, pretty much all leadership principles territory. Three questions, all of them the kind where you need a real story or you're dead in the water.
  • Describe a time you strongly disagreed with a senior decision. How did you push back, escalate appropriately, and eventually commit once the decision was made?
  • Give an example of how you raised the technical bar on your team, whether through mentoring, hiring, or design quality, and what measurable impact it had.
  • Tell me about a failure or incident you personally owned. How did you recover quickly, and what systems or processes did you put in place to prevent it from happening again?

“This one tripped me up a little because my instinct was to tell a story where I was right and the senior person came around.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

Jul 2026
Amazon ML Engineer coding round, got a classic interval merging problem dressed up in a conference room scheduling scenario. Pretty standard stuff but the edge cases around touching endpoints tripped me up a bit.
  • Given a list of potentially overlapping time intervals (as [start, end] pairs stored in arbitrary order), write a function that merges all overlapping or touching intervals and returns the smallest possible sorted list of non-overlapping intervals covering the same range.

“My first instinct was to just sort by start time and walk through, which is the right move.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

Jul 2026
Amazon ML Engineer interview that went deep into recommender systems. The whole session was basically one long question about collaborative filtering, and they kept pushing on the details until I ran out of confident answers.
  • Walk through how collaborative filtering works for recommendations, and compare user-based versus item-based approaches against matrix factorization. Cover both explicit and implicit feedback scenarios.
  • What are the tradeoffs between pointwise and pairwise ranking losses in a recommendation context? When would you pick one over the other?
  • How do you handle regularization in these models, and what goes wrong without it in large-scale recommendation systems?
  • Describe your approach to the cold-start problem for both new users and new items.
  • Popularity bias is a known issue in recommendation systems. How do you detect it and what techniques do you use to mitigate it?
  • How do you evaluate a recommendation model offline versus online? Walk through metrics like NDCG and MAP and explain what each captures.
  • How do you scale a collaborative filtering system to handle catalogs with millions of items and hundreds of millions of users?

“I started with the neighborhood methods and felt pretty solid there.”

View Post

Amazon·Machine Learning Engineer·Onsite - System Design / Architecture

Jul 2026
Amazon ML Engineer interview focused heavily on distributed training internals. Three questions, all technical, no fluff. The kind of round where you either know the mechanics or you don't.
  • Walk through the main parallelism strategies used when training models too large to fit on a single device. How does each one work, and what are the tradeoffs?
  • What are communication collectives like all-reduce, all-gather, reduce-scatter, and broadcast, and how do they actually show up in distributed training?
  • In tensor model parallelism, how do you split linear layers into column-parallel and row-parallel components? And what does it mean to alternate between column and row parallelism across layers, and why does that help?

“I started with data parallelism because it's the easiest to explain and bought myself time to think.”

View Post

Amazon·Machine Learning Engineer·Onsite - Behavioral / Leadership

Jul 2026
Behavioral round at Amazon for an ML Engineer role, basically one big question about decision-making under uncertainty. Pretty standard Amazon leadership stuff but the depth they expected was real.
  • Tell me about a time you made an important decision without having all the information you needed. Walk through the context, the options you weighed, the assumptions and risks you saw, how you gathered just enough data to move forward, and what happened. Looking back, what would you change and how did you handle the downside risk?

“This one is deceptively layered.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

Jul 2026
Amazon ML Engineer interview that went pretty deep into LLM fundamentals. Three big topic areas: Transformer internals, Mixture-of-Experts, and distributed training. More of a technical knowledge check than a coding round, which I wasn't fully expecting.
  • Walk me through the key components of the Transformer architecture and explain how self-attention works at a high level. Why does this architecture work better for language modeling than RNNs or LSTMs?
  • What problem does Mixture-of-Experts solve for large language models, and how does expert routing actually work? What are the main trade-offs you'd worry about?
  • Describe the different forms of parallelism used to train large models, explain what collective communication operations like all-reduce do, and give a concrete example of where all-reduce shows up when training a Transformer.

“I started with self-attention because I knew it well, explained the query-key-value thing, then talked about multi-head attention letting the model attend to multiple representation subspaces at once.”

View Post

Amazon·Machine Learning Engineer·Onsite - System Design / Architecture

Jul 2026
System design round at Amazon for an ML Engineer role, focused entirely on building a production RAG system for enterprise text. One long question that basically covered everything from ingestion to monitoring, which sounds manageable until you're actually in it trying to not forget half the components.
  • Design a retrieval-augmented generation system for enterprise text, covering the ingestion pipeline, retrieval strategy, prompt orchestration, grounding and citations, freshness handling, latency and throughput targets, privacy controls, hallucination reduction, evaluation approach, and production scaling and monitoring.

“This is basically a full ML systems design question crammed into one prompt.”

View Post

Amazon·Machine Learning Engineer·Technical Phone Screen

Jul 2026
Amazon ML Engineer interview, all technical fundamentals, no system design or behavioral that I can recall. It felt like a grad school exam more than a job interview, which I was not fully prepared for.
  • What is the bias-variance trade-off, and how do regularization techniques address it?
  • Derive the gradient of the loss function for logistic regression when L2 regularization is applied.
  • Compare ROC-AUC and PR-AUC. When should you prefer one over the other?
  • How do you detect and prevent data leakage when doing cross-validation?
  • What approaches exist for handling class imbalance, and how do they affect model calibration?
  • In what situations would you choose a tree-based model over a linear model?
  • How do you evaluate whether a model is well-calibrated, and what can you do to improve calibration?

“Bread and butter question but I still fumbled the explanation a bit.”

View Post

Amazon·Machine Learning Engineer·Onsite - Behavioral / Leadership

Jul 2026
Amazon MLE behavioral round, focused pretty hard on the ownership and cost-reduction angle. One question, but they clearly wanted a lot of depth on it.
  • Tell me about a time you found and fixed a problem that had a measurable cost impact, such as cloud spend, latency, operational burden, or labeling costs. What did you do, and what were the actual numbers before and after?

“This one has more layers than it looks.”

View Post