A portfolio should withstand the question “you or AI?”
A defense should not become a denial of AI use. A strong answer explains where your ownership was: framing, decisions, verification, debugging, trade-offs, and responsibility for boundaries.
A reviewer should see not only a polished result, but a path that can be reproduced or discussed.
Defense questions
- What problem did this slice address, and for whom?
- Which parts were AI-assisted, and which decisions were yours?
- Which signal checked the result, and what happened when the check failed?
- Which trade-off did you choose, and which alternative did you defer?
- What would you do next if you received new evidence?
Honest defense boundaries
- Do not present generated code as personal understanding without explaining verification.
- Do not invent contribution, metrics, user validation, interview feedback, or a job outcome.
- Name Unknown and propose a bounded next check.
- Keep evidence anchors, but do not publish secrets, personal data, or unnecessary operational logs.
Canonical lesson source · JavaRush
Open JavaRush material