I applied through a recruiter. The process took 3 months. I interviewed at Amazon in Nov 2021
Interview
A recruiter reached out to me on LinkedIn. Several friends shared experiences of low-ball attempts (especially in level). So, I was particular about my level (L5+ because my current role is L5), but the recruiter made me apply for a posting corresponding to L4, and promised that I will be considered for L5/L6. They finally offered me L4. The recruiter cited "feedback from the panel" as the reason for low-balling, but it was not difficult to see through their lies. And lastly, they claimed there's a sponsored "interview lunch", but they refused to reimburse my lunch even though it was well within the budget.
Rounds:
Coding screen followed by several rounds of interviews during virtual onsite
I decided to have all my virtual onsite rounds on a single day to avoid boredom
Interview questions [1]
Question 1
Typical leetcode questions. Even the recruiters advised me to "practice" on leetcode before the interviews. Total waste of time
ML depth round was not a depth round at all. Interviewer was busy taking notes and didn't probe the concepts further
"Name all the hyperparameters of boosting model - XGB and/or GBM"
Simpler alternative is to replace all interviewers with automated tests. I felt like I was talking to computers. I prefer leetcoding and attending multiple choice ML questions over talking to people who mechanistically follow instructions
Applied for Amazon AGI. After first round, it will go into full round of multiple interviews. Lots of modern LLM training technic questions. There are still some behavioral questions, but less than general Amazon roles.
Interviewed with 1 phone screen, 1 coding, 2 ml design and 2 lp rounds. Most questions were non-leetcode questions more related to day to day ml implementations. The questions were very practical.
I applied online. The process took 1 week. I interviewed at Amazon (Tokyo) in Apr 2026
Interview
The interview for the Applied Scientist position primarily focused on three core components: technical questions regarding machine learning, a live coding assessment, and a detailed review of my professional experience.