Your question is Scalable Training Infrastructure. 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 design a scalable training infrastructure for simulation and real-world teleoperation data for Figure AI?