Almost nothing you have shipped was shipped alone. That creates an awkward problem in interviews: say "we" throughout and your own contribution disappears; say "I" throughout and you sound like someone who takes credit for other people's work.
Experienced interviewers are listening for exactly this. Overclaiming is one of the fastest ways to lose a room, because it usually collapses under one follow-up question.
Use "we" for the outcome, "I" for the decision
This is the whole technique, and it resolves most of the difficulty.
"We moved the reporting pipeline off nightly batch onto a streaming model. I owned the backfill — specifically, how we replayed eighteen months of history without double-counting."
The team gets credit for the outcome. You get credit for a specific, defensible piece. Nobody has to guess where the boundary was.
Be concrete about your slice
"I worked on the migration" is unfalsifiable and therefore uninformative. The fix is to name the part you would still be able to answer questions about in two years.
Weak:
"I helped redesign the permissions system."
Better:
"I wrote the migration that mapped the old role table onto the new policy model. The tricky part was the accounts that had been grandfathered in and did not fit either schema."
The second version invites a good follow-up question. The first invites a sceptical one.
Say what you argued for, including when you lost
Decisions are more interesting than tasks, and disagreements are more interesting than decisions.
"I pushed for doing the backfill in place rather than standing up a parallel table. I lost that argument — the team was right that the rollback story was better with a parallel table. I would make the same call differently now."
This is disproportionately effective. It shows you have opinions, that you can hold them without ego, and that you update. Very few candidates volunteer a disagreement they lost.
Do not inflate scope, and do not deflate it either
Both failure modes are common, and the second is more common among people who are actually good.
If you were one of six engineers, say so. If the thing served a thousand users rather than a million, say a thousand. Precision reads as confidence. Vagueness about scale reads as hiding something, even when you are simply being modest.
Equally: if you did design the schema, say you designed the schema. Hedging your own work into "I was involved in" costs you credit you earned, and it makes the interviewer work harder to find out what you actually do.
The follow-up is the real test
Whatever you claim, assume the next question is "why did you do it that way?"
Claim only the things where you have an answer to that question. That single filter keeps you honest without shrinking you, because the work you genuinely did is precisely the work you can still explain.