Your question is Availability Under Heavy Load. 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 do you ensure your system remains available under heavy load at Liftoff?
Focus on how you would keep an ML-backed serving path reliable during traffic spikes, dependency failures, and degraded infrastructure. Explain the architecture choices, scaling strategy, and fallback behavior you would use.