Code review

Code review is a quality and knowledge-sharing gate: check correctness, maintainability, security, tests, observability, and product fit. Strong comments are specific, actionable, and tied to user or operational impact—not style debates without consequence.

| Review lens | What to look for | |-------------|------------------| | Correctness | Edge cases, race conditions, error handling | | Security | AuthZ, injection, secrets, SSRF | | Tests | Missing failure paths, flaky patterns | | Ops | Logs, metrics, feature-flag cleanup | | UX | Loading, errors, accessibility |

					❌ "This feels messy"
✅ "This N+1 query will timeout on accounts >10k seats; suggest batching or cursor"
				

On interviews: describe how you prioritize review comments, when you block merge, and how you give kind but direct feedback on risky changes.

Common pitfalls: nitpicking formatting while missing security; rubber-stamp reviews; bikeshedding names; no follow-up on unresolved threads.

The trade-off is review latency versus defects and knowledge silos reaching production.

Checklist:

• Comment on impact, not preference alone.
• Ask questions when intent is unclear.
• Verify tests cover new risk.
• Approve only when you would operate the change.