Your question is Real Time Vehicle Model 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).
You are deploying a perception model that must run on a vehicle in real time. You need to balance prediction quality with tight onboard compute and latency limits.
What techniques would you use for model optimization to ensure real-time performance on a vehicle?