This talk will present our research journey from energy-efficient RISC-V processor design toward neuromorphic computing architectures for embedded AI. It will highlight how principles from digital hardware design, processor architecture, and hardware/software co-design can be applied to build scalable, efficient neuromorphic systems.