This was a phone screen, consisting of three parts:
Values
Pretty standard behavioral questions, such as describing a situation.
Coding
Asked a hard LeetCode question about how to get the median of two sorted arrays (LeetCode #4).
Applied ML
Asked to explain an attention layer.
In the end, I feel like even for an MLE role, this was on the advanced side for both DSA and ML theory, which doesn't sound reasonable. On top of this, the interviewer had horrible communication skills and couldn't answer even the most basic questions properly.
Given two sorted arrays, nums1 and nums2, of size m and n respectively, return the median of the two sorted arrays. The overall run time complexity should be O(log(m+n)).
Explain the attention layer, self-attention, cross-attention, and multi-head attention layer.
The following metrics were computed from 18 interview experiences for the Snap Machine Learning Engineer role in United States.
Snap's interview process for their Machine Learning Engineer roles in the United States is extremely selective, failing the vast majority of engineers.
Candidates reported having negative feelings for Snap's Machine Learning Engineer interview process in United States.