Edge AI:
Maximum Power, Minimum Latency