Our aim is to protect services on devices as close as possible to where they are consumed. With the advent of machine learning it means we need to integrate cryptographic software to encrypt, decrypt and authenticate data onto other computing units than the main c of the Apple Silicon SoCs. This includes the Apple Neural Engine or the GPUs of the SoCs. Protecting machine learning models on device could be done through cryptographic computations using the computing units dedicated to ML.
You will join an extraordinary team, including world-class software engineers and cryptographers. We will partner you with a dedicated mentor who is an experienced member of our team. You will focus on implementing symmetric encryption algorithms to be run on the Apple Neural Engine and potentially the GPU on the Apple Silicon SoC. You will have to search for ideas to make the implementation as efficient as possible to take advantage of all the hardware optimizations available for inferences.