AIML - ML Engineer, Machine Learning Platform & Infrastructure
Job Description
Summary
At Apple, the AIML - On-Device Machine Learning group is responsible for accelerating the creation of amazing on-device ML experiences. Core ML is an example of an external facing product from this group. One important mission of the group centers around profiling, analyzing and optimizing ML inference performance across a remarkable variety of ML models, Apple devices and processors. These capabilities are already being used by several partner teams inside Apple (such as CPU, GPU, Neural Engine, speech understanding, Camera, Photos, VisionPro) as well as external App-Developers. If this role sounds exciting, we want to hear from you!
Description
This role provides a great opportunity to help scale and extend a significant ML benchmarking service used across Apple, in support of a range of devices from small wearables up to the largest Apple Silicon Macs. The role further offers a learning platform to dig into the latest research about on-device machine learning, an exciting ML front-tier ! Possible example areas include efficient inference, model compression, ML compilers, and/or federated learning.
Minimum Qualifications
- Strong ML fundamentals, experience developing code in one or more of training frameworks (such as PyTorch, TensorFlow or JAX)
- Experience with Transformers, CNNs or Stable Diffusion
- Understanding about performance modeling and profiling of computer systems, and how to optimize code efficiency for a given platform; Experience in computer system power analysis is a plus;
- A passion for edge / on-device ML; Experience with on-device ML frameworks such as CoreML, TFLite or ExecuTorch is preferred;
- An interest in software architecture, APIs, high performance extensible software and scalable software systems;
- Programming and software design skills (proficiency in Python and/or C/C++);
- Excellent collaboration and communication skills.
Preferred Qualifications
- Masters in Computer Science or relevant disciplines; PhD preferred.