John Thomson is an intern on the Deep Learning Algorithms team at NVIDIA. He’s currently in his third year of Computer Engineering at the University of Waterloo. His area of focus is optimizing LLM inference on structured workloads.