cswork / libraryOpen practice workspace

Interview experiences · Amazon

Amazon Intern Data Scientist Interview Experience — Two Back-to-Back Virtual Onsite Rounds

In progressOnsiteInterviewed 2026-04

This interview was purely a case of getting lucky. I got the interview invite on 4/12, then had two VO (virtual onsite) rounds back to back on 4/22, with no break in between. Both interviewers held PhDs. VO1 Standard LP (Leadership Principles) questions, and then they dug into a lot of detail based on my answers. Tell me about a most recent project Calculated risk A time you needed to extend a de…

This interview was purely a case of getting lucky. I got the interview invite on 4/12, then had two VO (virtual onsite) rounds back to back on 4/22, with no break in between. Both interviewers held PhDs.

VO1

Standard LP (Leadership Principles) questions, and then they dug into a lot of detail based on my answers.

  • Tell me about a most recent project
  • Calculated risk
  • A time you needed to extend a deadline
  • A time you needed to make a fast decision
  • How do you convince your teammates when they disagree with you
  • A time you led a team
  • A surprise obstacle or challenge
  • A time you needed to learn something totally new to you

They asked a lot of these, and I felt like the LP questions were the most important part.

Then we moved on to the technical part. At first they wanted me to write command line, but I wasn't very comfortable with it, so we switched to SQL instead. I did 3 very simple SQL problems. Basically if you know window functions and CTEs you can get through them — easy-to-medium difficulty. I hadn't done SQL before this, I crammed for it for one afternoon, and that was completely enough.

VO2

Lots of LP questions again, mostly similar to the previous round so I won't repeat them. They also asked:

  • Why Amazon
  • Why this position
  • Tell me about an NLP project you've done
  • Are you curious about anything in the NLP field? (roughly: while working on your NLP-related projects, was there anything you found particularly interesting or curious about — like something you thought needed to be solved or improved)
  • DDL
  • ML fundamentals (basic ML trivia questions)
  • How would you explain a confusion matrix to a group of high schoolers
  • Precision, recall, F1, AUC
  • When do you use precision vs. recall
  • How do you evaluate a model if it's multi-class
  • What is cross entropy
  • Have you ever run into a situation where manual evaluation was better than using an automated metric/function

That's roughly all of it, I don't remember the rest, but they were all fairly basic fundamentals questions too.

Both interviewers gave off a lot of positive energy, and the conversation was really pleasant. Not sure if I'll get an offer though.

Comments

Nothing here yet.

Join in on CSWork →