ASIC Design Engineer — Neural Engine DMA

Computational photography, live language translation, and countless other on-device AI capabilities depend on moving enormous amounts of data between memory and compute at exactly the right time. As part of our Silicon Engineering Group, you will design the high-performance DMA engines responsible for moving model weights and activation tensors between the memory subsystem and the Apple Neural Engine.

In this role, you will take end-to-end ownership of complex DMA subsystem blocks, shape technical direction across the team, and mentor engineers earlier in their careers. The work you drive ships in Apple products used by billions of people every day.

Description

As an ASIC Design Engineer on the ANE DMA Design team, you will drive the design of DMA subsystem blocks from microarchitecture specification through tapeout. ANE DMA's scale and complexity means grappling with:

decomposing tensor surfaces into efficient fetch sequences; maximizing throughput under real-world memory system conditions; maintaining data consistency across concurrent streams; multi-level arbitration and ordering correctness; sizing latency-tolerant buffering structures; and security and state isolation.