My application process consisted of a phone screen with a recruiter, two technical phone calls with engineers, and a 4-hour onsite interview.

Phone screen: Streamlined, effective, and easy to coordinate.

Technical call 1: 10 minutes of conversation about my experience and several real-time coding questions, with a preference for C++ only. They discouraged Python. Problems focused on data structures and algorithms.

Technical call 2: 10 minutes of conversation about my experience and several real-time coding questions, with a preference for C++ only. They discouraged Python. Problems focused on data structures and algorithms.

Onsite: Invited onsite and placed in a conference room with a whiteboard. One-hour time slots with several engineers from the team I was interviewing for, with each person asking several undergraduate-level whiteboard questions on the board. Topics ranged from C++, control, path planning, geometry, etc. In total, they asked me to solve 11 different questions on the whiteboard.

Takeaway: The interview process at Nuro is almost entirely paralleled to the technical problem-solving interviews at Google. They prioritize the ability to execute on entry-level computer science tasks (despite your experience level) and devalue soft skills or application-level knowledge. If you can’t solve software trivia questions or math problems that you haven’t seen since college, they’ll pass. Modern engineers use resources at their discretion to solve complex problems, yet they chose to analyze the ability for candidates to memorize software and algorithmic puzzles (how to tell if a polygon is closed, etc.). Give candidates a take-home project and let them demonstrate their ability to be a developer. Whiteboard problems only test a small fraction of a candidate's ability.

How to tell if a polygon is closed

What is a dot product

Build a matrix class in C++

Write a Lyapunov stability analysis for a first-order differential equation.