Senior AI-Driven GPU Performance Engineer

zymtrace is building an AI-driven agent that autonomously optimizes GPU workloads across PyTorch, CUDA, and hardware levels. You will contribute to the end-to-end execution path from user-space code to kernel behavior, using real telemetry to guide decisions.

You will work with a team of low-level engineers to generate actionable optimizations, validate with customers, and iterate rapidly to deliver measurable improvements in GPU efficiency.