Description
ZYPHRA is hiring a Research Engineer - AI Performance & Kernel Optimization to improve the performance of large-scale language model training and inference systems. The role focuses on developing and optimizing GPU kernels, tuning training and inference stacks, profiling bottlenecks, improving distributed training and inference, and optimizing performance across NVIDIA and non-NVIDIA accelerator platforms. Candidates should have strong systems and low-level performance intuition, experience with GPU kernel development and large-scale ML workloads, and familiarity with distributed training, parallelism, profiling, and hardware-software interactions.
