Gatik Machine Learning Engineer Interview Experience — A Point Cloud Config Design Question, Not LeetCode

A recruiter reached out to me on LinkedIn first. I talked to the company's HR, and they told me the company currently has fewer than 500 people, builds mid-range self-driving trucks, and just closed a Series D round, so they're hiring more.

Then I talked to the CTO. I mostly talked about what I'm currently working on, and my earlier PhD research direction matched what they do pretty well.

Then the first phone screen wasn't LeetCode. They asked a point cloud data preprocessing design question: you need to read a config from a YAML file, and that config needs to include a series of transformations, so you need to design:

• YAML/schema
• interface abstraction
• config validation
• dataflow modeling
• applying the transformations

I had assumed live coding meant LeetCode, so I was completely unprepared for this kind of design question, and I answered pretty haltingly. I asked the interviewer a lot of clarifying questions along the way, and the interviewer helped answer some of them.

I'm still waiting to hear back. I don't know if anyone else has had a similar experience — happy to discuss!