Member of Technical Staff, Kernel Engineering
Core
Write CUDA kernels and low-level optimizations to maximize performance of the vLLM AI inference engine on modern accelerators.
Role type
Senior IC kernel engineering specialist (GPU/Hardware)
Builds
High-performance inference engine kernels for vLLM
Domain
AI inference, GPU architecture, hardware optimization
Deliverable
production ML models
Required skills
CUDA kernel development, GPU architecture knowledge, C++, Python, performance profiling, benchmarking
Preferred skills
ML-specific kernel optimization, quantization techniques, multi-accelerator platform experience, compiler technologies
Technologies
CUDA, CuTeDSL, Triton, TileLang, Pallas, Nsight, rocprof, LLVM, MLIR, XLA
Responsibilities
Write and optimize kernels for NVIDIA GPUs and emerging silicon, integrate with hardware vendor teams, profile and benchmark performance
Seniority
Senior, hands-on IC
