GPU Engineer, Platform Architecture

GPU Engineer, Platform Architecture

Imagine what you could do here! At Apple, new ideas have a way of becoming extraordinary products, services, and customer experiences very quickly. Bring passion and dedication to your job and there's no telling what you could accomplish. Dynamic, inquisitive people and inspiring, innovative technologies are the norm here. The people who work here have reinvented entire industries with all Apple Hardware products. The same passion for innovation that goes into our products also applies to our practices strengthening our commitment to leave the world better than we found it!

Description

The Platform Architecture GPU group is looking for a talented GPU Engineer to join the Neural Accelerator effort with strong skills in performance analysis and development at the level of ML frameworks and lower-level kernel implementations.

Responsibilities

Analyze the performance of linear algebra and machine learning algorithms on Apple GPU platforms, pursuing investigations wherever they take you in Apple software. With our partner teams in Software Engineering and Hardware Technologies, formulate system-level strategies to address performance problems and unlock the next level of AI performance for our users. Work closely with MLX, MPS and CoreML teams on real ML use-cases in an end-to-end co-design effort, from early design exploration up to product launch.

Minimum Qualifications

BS degree Experience with software and hardware performance analysis and optimization Experience in GPU programming models such as Metal, CUDA, or similar Experience with ML frameworks, for example MLX, Pytorch, or similar

Preferred Qualifications

MS or PhD in Computer Science, Electrical Engineering, or equivalent 3+ years of relevant industry experience Experience working specifically in CUDA C++ on ML and/or linear algebra algorithms Experience optimizing LLM inference for low latency at the implementation level Experience optimizing LLM inference at scale in the cloud or datacenter Ability to communicate across both hardware and software organizations