with ML software and hardware engineers to understand how models execute and where performance and efficiency can be improved.The work could include ML compilation, PyTorch integration, runtimes and drivers, performance analysis or embedded software. There will also be opportunities to contribute to technical design, take ownership of larger engineering challenges … using source control, code review, testing and continuous integration.Experience across every part of the ML software stack isn't expected. Knowledge of ML compilers, PyTorch internals, runtimes, drivers, embedded software, performance optimisation or hardware acceleration would be valuable, but we encourage applications from people whose experience doesn't cover every ...