
Amazon Data Scientist Interview Questions
& Process
Real candidates share what happened, how many rounds they had,
and how the experience turned out.
Based on 246 interview experiences · FREE TO READ
Candidate interview experiences
First-hand accounts from people who interviewed at Amazon.
Applied Scientist
I got a referral to the job because of my current collaboration with Amazon for my dissertation. I'm a 2nd-year PhD student in Computer Science. The first part was an online assessment (OA) which mainly had Leetcode easy and medium questions and the leadership principle test. The next phase was the technical interview, where I was interviewed in 2 different sessions, each session with an Applied Scientist for 1 hour. The interview mainly focused on my research collaboration with Amazon and my previous research works during undergrad.
- What was your research mainly about?
Applied Scientist
Online screen with a coding exercise and behavioral questions. Then onsite loop interviews with presentations, different coding exercises (including DFS), an ML system design question, depth and breadth knowledge questions, and a lot of behavioral questions.
- Design a Github Copilot-style agent.
Applied Scientist
Recruiter contacted me via LinkedIn. I was preparing more on resume and leadership principles but the screen round was purely statistical. Screen included: -Bayes theorem question -Can we fit linear regression on 100 rows, 1000 columns -var(x) = 2var(y), with the same expectation, what is the best possible way to get weight measurement -Daily rainfall data - how to get the rainy season time
- -Bayes theorem type question
- -Can we apply linear regression when we have 100 rows and 1000 columns
- -var(x) = 2var(y), with same expectation, what is the best way to do weight measurement
Amazon Data Scientist Interview Questions
Quoted word for word from Amazon interview reports.
“Given a PDF and a random variable, draw some specific number of points from it using a given list of probabilities”
Read reports →“Given a time limit and edge costs, check if the destination can be reached within that limit. If it can, return the minimum time and cost; otherwise, say the destination can’t be reached within the limit.”
Read reports →“What is KL divergence and how does it connect to cross entropy loss in practical use?”
Read reports →“How is linear regression different from logistic regression, and what kinds of loss functions are used in each case?”
Read reports →“What is the closed-form solution to linear regression, and when is it preferable to use gradient descent instead? Also, what is the time complexity of matrix inversion?”
Read report →“What is the role of dropout and what is the role of the activation function?”
Read report →“Given a graph and specific origin and destination nodes, how would you determine a shortest path from the origin to the destination?”
Read report →“Can you build an anomaly detection tool using pandas? How would you compute a rolling average in pandas?”
Read report →“Given these distribution centres and the costs linked with each route, design an algorithm to determine the best overall strategy.”
Read report →Formats, difficulty and experience
Across all 246 Amazon interview reports.