Jobs

Inferact is seeking a Member of Technical Staff, Kernel Engineering to join their team and optimize the vLLM AI inference engine. This performance engineer will be responsible for writing kernels and low-level optimizations to maximize performance across various accelerator types, from NVIDIA GPUs to emerging silicon. The role requires deep experience with CUDA kernels, a strong understanding of GPU architecture, and proficiency in C++ and Python for high-performance coding. The ideal candidate will be obsessed with benchmarks and squeezing every percentage point of speedup to make vLLM the fastest inference engine globally.