Khubaib is a senior deep learning performance architect at NVIDIA, dedicated to maximizing GPU efficiency and shaping next-generation architectures. His recent work involves evaluating inference performance of large language models and guiding software teams in optimization efforts for high-performance MLPerf Inference submissions. Previously, he worked on performance modeling and simulation of deep learning models to inform GPU architecture design. Before NVIDIA, he contributed to CPU architecture research and development at Apple and Intel. Khubaib holds a master’s degree and PhD in Electrical and Computer Engineering from the University of Texas at Austin.