Your question is 10ms Edge Inference Optimization. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How would you optimize a PyTorch model to ensure it meets a strict 10ms inference latency budget on edge hardware for Figure AI robots?