GPU Kernel Engineer
Core
Design, implement, and optimize custom GPU kernels for large-scale LLM training and inference workloads.
Role type
Senior IC GPU Kernel Engineer
Builds
Optimized GPU kernels integrated into PyTorch, JAX, and internal runtimes
Domain
High-performance computing + Large Language Models
Deliverable
production ML models
Required skills
C++, CUDA, ROCm, PTX, Triton, JAX Pallas, GPU memory models, performance profiling, low-level GPU execution
Preferred skills
AMD GPU optimization, JAX FFI, vLLM, TensorRT, TPUs, XLA, accelerator programming, open-source ML systems contributions
Technologies
PyTorch, JAX, Triton, CUDA, ROCm, PTX, JAX Pallas, vLLM, TensorRT
Responsibilities
Design and optimize custom GPU kernels; profile end-to-end ML operation performance; integrate low-level kernels into frameworks; develop performance models and identify bottlenecks; collaborate with ML researchers and hardware vendors; contribute to tooling and testing frameworks
Seniority
Senior, hands-on IC
