Kernel Engineer: AI Kernels for Long Context with Equity

Magic is hiring a Kernel Engineer in San Francisco to design, implement, and optimize high-performance kernels for long-context training and inference. You will tackle memory usage, data movement, and throughput challenges in real-time workloads. You'll work across training, inference, and RL teams, focusing on robustness, extensive testing, and quality. Experience with AI accelerators and GPU kernel frameworks is essential; visa sponsorship is available in some cases. #J-18808-Ljbffr