Zoox

Software Engineer, Planning & Control, 2018-2021


I joined Zoox when the Planning & Control team was less than 20 people. While there, I led multiple projects that required coordination with several other teams across hardware and software.

Major accomplishments include:
  • Led the design and development of the software responsible for calculating and enforcing lateral and longitudinal motion limits that can be dynamically updated depending on the operational domain and configuration of the vehicle. A major challenge with this project was that in needed to work for both two-wheel and four-wheel steering vehicles, requiring the invention of a completely new approach to limits. I implemented this software throughout the software stack, from top-level planning algorithms down to the lowest level, safety-critical control systems. This involved orchestrating major architectural changes to various software components to support the dynamic nature of the limits.
  • Reduced path tracking error of our level 3 testing fleet by ~50% by developing a novel low level steering control algorithm.
  • Developed a model that accurately simulated the level 3 vehicles' low level control behavior, making our tests suites more reflective of real world behavior.
  • Discovered a critical issue with the Prius C testing fleet and presented my detailed findings to the executive team, leading to the decision to scrap the Prius C fleet and return to using Highlanders as the primary test vehicle.
  • Was selected for a task force to re-design the general algorithms interviews for prospective Zoox software engineers, and personally designed the controls algorithms interview for the RMC team.