Code review

Code review evaluates correctness, maintainability, security, tests, observability, and integration risk. Git supplies the diff and discussion thread; the reviewer must reason about runtime behavior and operational impact.

Effective review layers:

1. **Behavior** — does it solve the right problem safely?
2. **Tests** — are risky paths covered?
3. **API and data contracts** — migrations, backwards compatibility
4. **Style** — ideally automated via lint/format tools

On interviews: review scope, respectful feedback, risk-based depth, ownership rotation, and how automation reduces repetitive load.

Common pitfalls: style-only review missing logic bugs; blocking on subjective preferences not encoded in tooling; rubber-stamping large diffs under time pressure.

The trade-off is review rigor versus team velocity — encode mechanical rules in CI, reserve humans for judgment.

Checklist:

• Review behavior and risk before nitpicks.
• Ask for tests around failure modes.
• Use tooling for formatting and lint rules.
• Keep feedback actionable and kind.