Principal Engineer, CUDA UMD - GPU Kernel Scheduling
Core
Architecting and implementing CUDA driver features to optimize GPU kernel scheduling for AI/ML workloads.
Role type
Principal Engineer, CUDA UMD (GPU Kernel Scheduling)
Builds
CUDA platform features and improvements for deep learning, scientific computation, and self-driving cars.
Domain
GPU system software, parallel computing, CUDA programming model
Deliverable
production ML models | product features
Required skills
C/C++ programming, CUDA driver development, OS interfaces (threads, process control, virtual memory), multithreaded programming, large codebase management, cross-team coordination
Preferred skills
Parallel computing (CUDA programs/libraries), system-level architecture (interconnects, memory hierarchy), memory coherence models, kernel mode development, Linux systems software
Technologies
CUDA, Linux, C, C++
Sourced via workday · Listed on CareerPlan, which tracks 70,000+ jobs from 20+ sources.