take home data science homework is free form, you are given a pipeline with a bunch of classifiers that'll predict whether a user is retained or not based on swiping activity, you're told to add whatever you want to the pipeline to showcase your skills the tools are the traditional sklearn/numpy/matplotlib etc and the algos involved are regression/xbgoost/other voting algos
Had an HR round and then a technical round where we talked about projects and then a case study. The issue is that the case study is a tough problem that they are dealing with; that is, something they have a whole team dedicated to and have been thinking about for months to years. Yet you are supposed to solve it in just a few minutes! So it really is not a fair assessment but that's how all tech interviews are today. One thing was the interviewer kept interrupting and didn't let me finish so I just figured he was trying to drive me toward to a set of exploratory questions on the case study. I did answer every question but got turned down.
discussion of semantic segmentation, describe Mask RCNNs/YOLO algorithms, cross entropy, graph/DFS type problems, design an optimizer for any cost function f
One main question was related to my motivation for joining the company and choosing this position.
No difficult questions. Standard
No particularly difficult questions,
How do you implement feature selection in text classification?
Why do you think you are the best candidate for this position?
What did i do in the last project?
Went over resume/elaborated on specific details that apply to the job.
Viewing 2031 - 2040 interview questions