Your model provider is down. How do you keep the product usable instead of showing an error?
A provider outage should cost you quality, not the feature. The answer that scores is a pre-decided fallback ladder, a timeout policy that treats slow as down, and the discipline to test the path you hope never runs.
Updated Sep 2026 · Grounded in real GenAI, LLM, and AI/ML engineering interview loops and written to a senior-engineer editorial bar.
A provider outage should cost you quality, not the feature. The answer that scores is a pre-decided fallback ladder, a timeout policy that treats slow as down, and the discipline to test the path you hope never runs.
Lead with where the obvious approach breaks, because that is the judgment they are screening for — most candidates jump straight to the happy path and lose the room.
Then walk the failure back through the pipeline in order, naming the one metric the customer's exec sponsor actually cares about before you propose the fix.
