The role involves the C++ code development and maintaining of GPU performance models from the shader core up to the full system. You will collaborate with experienced architects, compiler engineers, RTL designers and software teams to assess the feasibility of new hardware and software features, running performance experiments and reporting them. You will work on the analysis of existing and new workloads to identify bottlenecks in the design or software and working on solutions with the team. You’ll be architecting new Metal APIs, simulator features, and brainstorming novel ideas to boost simulation performance and accuracy.