I applied online. The process took 5 days. I interviewed at comma.ai (San Diego, CA) in Jun 2021
Interview
Software engineering challenges and technical questions. Both concrete questions of machine learning and brain-teaser questions. Concrete questions about self-driving and software engineering too. Questions likely depend on specific position applied for.
Interview questions [1]
Question 1
How to build a vision system to detect motion from video.
Highly suspicious interview process. First, you do an extensive coding project; pick a task from their product's code repository bounty board and complete it in exchange for a small amount of money. Keep in mind, it's a codebase of thousands of lines that you have to learn and understand before starting a task. They also openly admit that your code written during the interview process will ship in their product. Then they evaluate you to see if they want to hire you on as an "intern" in a trial period, during which it was unclear if you got paid at all. Several months later, they re-evaluate you for a promotion to "full-time." Although the person I spoke to was civil, they were quite dodgy to questions and eventually stopped replying at all. Other people's mileage may vary, but somehow, I felt the whole experience was even worse than a Leetcode interview.
Interview questions [1]
Question 1
Please pick a task from our bounty board and complete it. After you are done, email us and we will schedule a call to evaluate you as a possible intern for our company. All people in our company start off as interns and eventually get re-evaluated for a full-time position in several months.
I applied through other source. I interviewed at comma.ai in Jan 2021
Interview
The interview involved a phone interview and online programming hacker rank challenge. The interview over the phone was very straight forward, they are looking for top programmers in the world, so it is a very difficult challenge.
Interview questions [1]
Question 1
The phone section of the interview was only screening.