Waymo's Machine Learning Engineer Interview: Navigating Ego Behavior Models and Coding Challenges

Software Engineering
,
1

Waymo | Machine Learning Engineer | Interview Experience

Interview Date: Not specified
Result: Rejected
Difficulty: Medium

Interview Process

  1. In July, HR reached out proactively without an application, and I spoke with the manager. There was a good match, but due to my graduation timeline, the interview was scheduled for September.
  2. The first round was a system design interview focused on an ego behavior prediction model, which was relevant to the position. The questions were detailed, covering inputs and outputs, what outputs were desired, and how to measure uncertainty.
  3. I was quickly notified to move forward to the final two interviews.
  4. The coding interview involved writing a linear equation to find the optimal solution (similar to gradient backpropagation). The interviewer was unengaged, changing the interview time last minute and providing no feedback during the coding process, which made for a negative experience.
  5. The second coding interview involved implementing a basic Transformer model, which was fairly standard and not particularly difficult.
  6. After a week, I received a rejection call from HR.

Technical Questions

  1. Design an ego behavior prediction model for autonomous vehicles (Machine Learning, Autonomous Driving, Uncertainty)
  2. Find the optimal solution for a linear equation (Mathematics, Search, Brute Force)
  3. Implement a basic Transformer model (NLP, Deep Learning, Transformer)

Tips & Insights

I have interviewed with Waymo four times, including for internships and full-time positions. Each experience had its challenges, often involving questions related to autonomous driving scenarios. The overall interview difficulty is medium, but the questions tend to be closely tied to real-world applications in autonomous driving.