you prepare one version of yourself for an interview. then a different one walks in.
the prepared one is cleaner. right sequence of projects, clear explanation of what you are building, answers that sound finished. it is sort of the version that has already decided how the story should land. most of the work before the call is just spent protecting that version.
then the conversation starts and something else appears. sometimes sharper, sometimes more uncertain, sometimes just more honest about what is still half formed. the prepared version was a plan. the one that shows up is closer to the actual state of your thinking that day, i think.
i have started paying attention to that gap. not as a failure of preparation, but as information. the distance between the story you meant to tell and the one that actually came out usually says more than either story on its own.
part of it is just nerves, right. under pressure the polished edges fall off and you default to how you actually talk about the work. part of it is the other person. good interviewers do not let you stay inside your prepared arc for long. they pull on a detail, or stay quiet after an answer that was supposed to close the topic. the prepared version is brittle in those moments. the live one has to improvise.
i notice this most when i talk about the ai manager stuff. the clean version sounds like a product thesis. the version that shows up under questions is messier. it includes the parts that still do not work, the assumptions still being tested. that messier version is closer to the truth of the build. it is also harder to control.
there is a temptation to close the gap. prepare harder until the live version matches the planned one. i am less sure that is the right goal. when the two diverge a lot, it usually means the story you are telling yourself about the work has drifted from the work itself.
the useful interviews have been the ones where that drift became visible. not because i performed well, but because i left clearer on which parts of my own narrative were load bearing and which were just habit.
anyway. i still prepare. i still write down the arc. i just trust it less as a prediction of what will actually be in the room.