Notes on “Machine Learning System Design by Chip Huyen” (for Interviews)

Photo by Nagara Oyodo on Unsplash

These are my notes on this article by Chip Huyen. The article talks about some factors that are involved in designing machine learning systems, and what to watch out for in interviews on the same.

As a note, I found Chip’s book Designing Machine Learning Systems an excellent resource for anyone starting or willing to improve their skills in this field. I would recommend a read. (This is not sponsored by any means.)

Interviews

• The major issue with Machine Learning interviews seems to be the lack of a standard criterion by which to judge a candidate. This makes a lot of sense considering how varied the requirements of each project are.
• Interviewers generally look for what they are familiar with and often this means that an ideal candidate would be someone who thinks along similar lines. This is especially interesting in ML because there are an infinite number of ways to approach a problem.
• It would be helpful for the candidate to understand what kind of answers the company would be looking for based on their previous work.

Compute requirements

• Contrary to the vast amount of research in the ML space on improving models and focusing on metrics, in…