High-performance ML workloads on Arm CPUs requires the co-development of algorithms and highly optimized CPU kernels. In CT-ML (Central Technology, Machine Learning), rapid kernel prototyping is crucial for exploring algorithms and assessing trade-offs between model accuracy and performance. Successful prototypes are essential to drive future CPU architecture development and also deliverables to Central Engineering for final production.
This position is part of a dedicated team within the CT-ML group to focus on analyzing ML workload, rapid prototyping of highly optimized CPU kernels to drive model performance and accuracies.
- Strong interest and passion for implementing high-performance kernel code in a dynamic environment.
- 4+ years experience in implementing high performance CPU kernel with vector and matrix extensions.
- Experience measuring and understanding performance
- Experience in creating an efficient kernel code development framework including tools and testing
- Deep understanding on CPU architecture
- Knowledge of ML models and algorithm is a plus
- Advanced degree or equivalent experience in Computer Architecture and Software are a plus
Arm is committed to global talent acquisition, offering an attractive relocation package. With offices around the world, Arm is a diverse organization of dedicated, creative and highly talented engineers. By enabling a dynamic, inclusive, meritocratic, and open workplace, where all our people can grow and succeed, we encourage our people to share their unrivaled contributions to Arm's success in the global marketplace.
