3 tech rounds: 2 coding with IC and 1 chatting about research experience with HM. Coding is easy to medium and mainly machine learning focus. One is about sampling and one is about gradient descent.
Interview questions [1]
Question 1
Implement Inverse CDF sampling; find the square root of the given number using gradient descent.
Very good. Interviewer is very friendly and respects the candidate. Questions are thoughtful. Expect to be very familiar with modern deep learning architectures and being able to code from scratch.
Interview questions [1]
Question 1
How to implement a Transformer from scratch in PyTorch
three technical interviews focused on problem-solving, coding, and core computer science fundamentals. Depending on the role, there may also be discussions around past research or project experience, system design, or domain-specific knowledge
Round of introductions. First discussion about research with a focus on current project and past experience, then worked on leetcode-style coding questions with subquestions that talk a bit about system design.
Interview questions [1]
Question 1
How to implement a hash-based aggregation in Python? (Extend to out of memory, extend to accelerators)