What state should persist between agent turns?

Example Answer

The way I'd think about it is this: Only the state that is needed to continue the job reliably should persist. That usually includes user intent, confirmed constraints, completed actions, pending approvals, important tool outputs, and identifiers needed to resume safely.

I do not want the agent carrying forward every conversational detail just because it might be useful. Excess state makes prompts noisy and increases the chance of stale or irrelevant context steering the next turn.

I also separate durable workflow state from ephemeral reasoning. The durable state should be explicit, inspectable, and safe to reload after interruption. The model’s scratchpad should not become the system of record.

What matters in an interview is not only knowing the definition, but being able to connect it back to how it changes modeling, evaluation, or deployment decisions in practice.

Common Poor Answer

A weak answer is saying you should persist as much context as possible so the agent stays smart. More state often makes agents less reliable, not more.