Responsibilities

• Profile and improve the performance of ML workloads on different platforms (e.g. Nvidia, Apple, Qualcomm)
• Build highly-optimized GPU kernels for our Inference Engine
• Distill highly technical project outcomes in layman-approachable technical blogs to our customers and developers
• Mentor junior wizards and interns

Qualifications

• Skilled in debugging, profiling, and performance hillclimbing for GPU kernels
• Expert knowledge of parallel programming
• Familiarity with Metal and/or CUDA/Triton
• Advanced understanding of modern Deep Learning workload characteristics
• Basic understanding of fundamental Machine Learning concepts
• Android/Windows experience is a plus