I applied through a recruiter. The process took 4 weeks. I interviewed at Amazon (Seattle, WA) in Apr 2025
Interview
The interview process began with a technical phone screen with the hiring manager. We mostly dived deep into one of my projects on my resume and ended with a Leetcode (medium) question on DP from Blind 150. Then I had 6 rounds of interviews for the virtual on-site. Since I am a PhD student, one of the rounds was a tech talk followed by Q&A, and the other was the Bar Raiser, which only focused on Leadership Principles. All other rounds were technical, focusing on science breadth, depth, and coding. You should know your fundamentals for different ML algorithms, Transformers, and current GenAI techniques (pre-training, fine-tuning, post-training, prompting, etc.)
Interview questions [1]
Question 1
Explain attention mechanism in Transformers. How is BERT different from GPT? Differences between DPO and RLHF
I applied through a recruiter. The process took 3 weeks. I interviewed at Amazon (Seattle, WA) in Feb 2025
Interview
Weird interview. Only the phone screen. Encountered a weird interviewer. Asked questions and about papers not related to my research at all. No fundamental ML or otherwise questions asked. The interviewer couldn't understand the mathematics I was explaining and had a very weak understanding of ML and CV concepts.
Interview questions [1]
Question 1
Explain different kinds of transformer architectures.
The interview process started with a behavioral phone screen interview. It consists of questions that need to be answered using the STAR method. The next step would have been the full loop but I got rejected earlier.
Interview questions [1]
Question 1
Describe a difficult conversation you had with your manager