Machine Learning Architect, Platform Architecture

What you'll do

• Create optimized implementations of ML workloads on Apple silicon including the Neural Engine, GPU, and CPU.
• Develop performance models and simulations of future hardware in collaboration with IP and SoC architecture teams.
• Create high-level performance models of emerging ML techniques to analyze system architecture trade-offs.
• Evaluate ML model optimization techniques from literature through experimentation and analysis.
• Propose new ideas based on research to inform both hardware and algorithm directions.
• Provide guidance to software and algorithm teams on how to run workloads efficiently on SoCs.
• Prototype algorithms on CPU/GPU/Neural Engine while analyzing performance metrics and creating complexity models.

What we're looking for

• Bachelor's degree required; MS or PhD in EE, CE, CS, or a related field is preferred.
• Experience with C/C++ and/or Python programming languages.
• Experience with hardware IPs including ML accelerators, GPU, CPU, or image/video processors.
• Experience with ML frameworks like PyTorch and implementing machine learning algorithms efficiently.
• 20+ years of relevant experience.
• Experience optimizing and deploying ML models or runtime frameworks in production inference and training environments.
• Ability to prototype algorithms on CPU, GPU, and Neural Engine while analyzing performance metrics and creating complexity models.
• Understanding of compilers and strong verbal/written communication skills for cross-team collaboration.