Define product features and own the architecture for compute, memory, interconnect, and high-speed interface subsystems in AI inference chips.
Must-have
Nice-to-have
Not specified