JOBSEARCHER

Software Engineer - Planner GPU Compute

ZooxSomerville, MAL6 LeadOctober 26th, 2026
Overview: In this role, you will optimize the motion planner's GPU-based workloads, improving latency and resource efficiency across architectures. You will work with the Planner Compute team to identify hotspots, tune CUDA kernels, and adapt the planner for different GPU generations. You'll optimize machine learning inference paths and serve as a CUDA expertise resource for engineering teams. The role offers impact on safety-critical autonomous driving software and performance at scale, across cross-functional collaborations. Compensation / Benefitspaid time off Zoox Stock Appreciation RightsAmazon RSUshealth insurancedisability insurancelife insurance
Responsibilities:
  • Analyze performance metrics to identify GPU hotspots and optimizations
  • Adapt planner to multiple GPU architectures with varying resources
  • Optimize ML models for latency and GPU memory usage
  • Provide CUDA & GPU performance expertise to engineers in Planner
Key requirements:
  • BS in computer science or related field 6+ years of experience
  • Strong knowledge of CUDA for recent GPU microarchitectures (e.g., Ampere, Blackwell) and debugging/optimizing GPU kernels with NsightStrong knowledge of C++ and experience in large code bases, Linux development environments
  • Experience in development, debugging, and profiling of complex multiprocess systems (robotic systems, game engines)CUDANsightC++LinuxGPU kernel developmentML model optimization