SW ML Optimization Engineer

At Apple, our Platform Architecture group is responsible for connecting our hardware and software into one unified system. You'll collaborate with engineers across Apple to design how all of our technologies work in unison, drive development of our renowned system-on-a-chip architecture and develop forward-looking prototype systems and software. Our team is driving performance enhancements in application and system software and developing novel algorithms to deliver integrated, highly optimized solutions based on Apple Silicon.

In this role, you will analyze existing and new workloads to identify performance bottlenecks in the hardware and/or software. Working with your colleagues, you will address performance limitations and provide recommendations for Apple hardware and software improvements. In addition to working directly with developers, you will identify patterns of performance challenges on Apple silicon, emerging new usage models, and provide feedback to the silicon and software teams for potential improvements.Profile and analyze AI/ML workloads across Apple Silicon compute engines (GPU, ANE, and CPU) to help identify performance bottlenecks, working alongside senior engineers. Build targeted microbenchmarks to characterize the strengths, weaknesses, and usage patterns of different devices and workloads. Develop and tune optimized implementations of software workloads on Apple Silicon (GPU and CPU), learning to apply the latest instruction sets and frameworks. Contribute to the development of forward-looking AI/ML workloads that represent where the industry is headed. Partner with system and tools teams to prototype performance experiments and help build performance analysis instruments, libraries, and frameworks. Grow toward a system-wide understanding of the stack-from high-level APIs down to the hardware.Bachelors degree in Computer Science, Computer Engineering, Mathematics, Electrical Engineering, or a related quantitative field (or equivalent practical experience). Experience with GPU or parallel programming-e.g., Metal, OpenCL, CUDA, or similar-through coursework, personal projects, internships, or research. Experience with profiling/performance analysis tools (e.g., Xcode Instruments, VTune, Nsight Compute, or equivalent) and basic performance analysis concepts. Development experience in Python, C or C++.Solid foundation in mathematics, algorithms, and/or computer architecture fundamentals. Experience writing or tuning compute kernels (e.g., GEMM, attention, or other numerically intensive routines). Exposure to ML frameworks such as PyTorch, and to AI/ML, graphics, or HPC workloads and benchmarks. Coursework or projects involving parallel computing, numerical methods, signal processing, or performance optimization. Interest in (or exposure to) the deeper stack - drivers, firmware, compilers, or low-level libraries. Interest in Apple Silicon and its frameworks (Metal, MLX, Core ML). Curiosity about hardware/software co-design and a demonstrated drive to learn independently. Strong communication skills and the ability to collaborate effectively across teams.