1st onsite interview: Technical/Research questions included: YOLO and its derivatives, SVMs and kernels, Metric learning, Segmentation (semantic vs instance), non-dl image matching (i.e SIFT features), clustering methods, supervised vs non-supervised learning etc.
4th onsite interview: Technical question How would you improve amazon's core services using computer vision techniques. The question actually overlapped with the discussion with the 2nd interviewer, so I had to come up with a different novel idea. Coding question For a list of rectangles/subarrays find the average of each one on a 2D Matrix
5th onsite interview: Coding question Find the second largest value in a binary tree
Two interviews: The first interview was on my research work and one coding question (write code for a basic text tokenizer). The second interview was on an ML design question (How will you go about designing a Machine Translation system) + behavioral question (Tell me about a time when you had limited time to work)
Q: What's the project you most proud of?
tell me about a challenging project that you have done.
What is self-attention and how does a transformer work?
Questions were on EM algorithm, Ensemble techniques, and ML infrastructure pipelines on the technical side. Two Leetcode Medium questions.
List a technical challenge & how you overcame it.
The question asked was a Leetcode deserialize serialize the tree.
Viewing 1081 - 1090 interview questions