Description
NVIDIA is hiring a Senior AI Training Performance Architect to analyze, profile, and optimize AI training workloads across GPU hardware, drivers, deep learning frameworks, and proprietary simulators. The role focuses on identifying performance bottlenecks, implementing production-quality software, supporting MLPerf Training benchmarks, developing automation tools, and contributing to hardware and software roadmaps. Candidates need a PhD or equivalent experience with 5+ years, or an MS with 8+ years, along with deep learning, computer and GPU architecture, performance modeling, and C++, Python, and CUDA expertise. The base salary range is 184,000 USD to 287,500 USD for Level 4 and 224,000 USD to 356,500 USD for Level 5, with equity and benefits; applications were accepted until July 28, 2026.
