Shopify Engineering Manager Interview Experience
Shopify · Staff
Interview process
didn't go well: (1) missing image data handling was not a good answer. i realized a better solution right after (2) hierarchical classification, how to do level by level? (3) overall infrastructure design including model training infra and streaming data serving
Interview rounds · 2
- 1
Recruiter screen
Behavioral- Q1. Tell me about yourself.
- Q2. talk about a project they are working on with technical blog shared
- 2
Technical round
System Design- Q1. Classify product into hierarchical classes and generate attributes for products. There are 5k leaf classes and assume every product can be classified into a leaf class
- Q2. I designed a deep learning based classifier using both text data (product tile, description etc) and image. I use a bert model to encoder product text information and an image encoder (vision transformer, CNN) to encode image and then concatenate them together and sent to a FNN for classification. Follow up question include: (1)5k classes? is that too many classes? I modified to do classification level by level since it is hierarchical. (2) what about missing image? (3) how to store the result? (4) I also briefly talk about using generative model and class label to generate attributes
Tips from the candidate
Shopify has technical blogs. I would say read about their technical blogs and ask yourself some follow up questions
Company culture
recruiting team is professional. overall is a friendly interview