Machine Learning Performance Engineer

The role involves optimising the performance of machine learning models, including training and inference, through a whole-systems approach involving GPUs, storage, and networking. Candidates should possess knowledge of low-level GPU programming, debugging tools, and relevant libraries such as CUDA, Triton, and cuDNN. Experience with distributed GPU training, networking technologies like NVLink, and system-level optimisation is essential. The position requires a curious mindset and problem-solving skills, with fluency in English. It offers a chance to work on innovative problems in a dynamic trading environment without a fixed skill set.