Research Engineer - AI Performance & Kernels
Core
Develop and optimize high-performance GPU kernels and distributed training systems for large-scale language-model pretraining and inference.
Role type
Senior IC research engineer (AI performance & kernels)
Builds
Optimized training and inference stacks for large mixture-of-experts models
Domain
AI / High-Performance Computing (HPC)
Deliverable
production ML models
Required skills
GPU kernel development (PTX, CUDA, HIP, Triton), distributed training optimization, hardware-software interaction analysis, memory hierarchy optimization, profiling and debugging, parallel computing (data/tensor/pipeline parallelism), non-NVIDIA accelerator experience (AMD, AWS Trainium, Google TPU, ARM, Intel, ASICs)
Preferred skills
HPC experience, background in physics/math/theoretical CS/EE
Technologies
PTX, CUDA, HIP, Triton, AMD MI300X, MI355X, AWS Trainium, Google TPU, ARM, Intel, collective communication libraries
Responsibilities
Develop and optimize kernels for large-scale machine-learning workloads; Tune performance across training and inference systems on GPUs and other accelerators; Profile systems and remove bottlenecks in memory movement, communication, scheduling, and compute utilization; Optimize distributed training and inference for large mixture-of-experts models; Improve portability and performance on non-NVIDIA hardware; Partner with research and infrastructure teams to translate systems improvements into measurable gains
