Эта вакансия в архиве
Посмотреть похожие вакансии ↓Описание вакансии
Текст:
TL;DR
Software Engineer (Planner GPU Compute) (Autonomy Software): Optimizing GPU-based algorithms for a motion planner in a commercial robotaxi system with an accent on CUDA performance, ML inference latency, and GPU memory usage. Focus on profiling GPU hotspots, adapting the planner to multiple GPU architectures, and debugging complex multiprocess systems.
Location: Boston, MA or Foster City, CA; hybrid workplace
Company
Develops purpose-built commercial robotaxis combining self-driving hardware and software.
What you will do
- Analyze performance metrics to identify GPU hotspots and optimization opportunities.
- Adapt the motion planner to multiple GPU architectures with different resource profiles.
- Optimize machine learning models for inference latency and GPU memory usage.
- Support Planner engineers as a subject matter expert in CUDA and GPU performance.
Requirements
- Bachelor’s degree in computer science or a related field.
- 6+ years of professional experience.
- Strong knowledge of CUDA and recent GPU microarchitectures, including Ampere and Blackwell.
- Experience debugging and optimizing GPU kernels with tools such as Nsight.
- Strong C++ skills and experience working in large codebases in Linux environments.
- Experience developing, debugging, and profiling complex multiprocess systems such as robotic systems or game engines.
Culture & Benefits
- Hybrid work environment.
- Work on performance-critical autonomy software for commercial robotaxis.
Музыка
Осень
Поддержка