01 Read code before you trust it
“Critically reviewed and improved AI-generated code” (opens the source)
A plausible patch can still break a valid input. Trace it.
Software interviews when AI can write the code
AI can draft the patch. You still need to explain the bug, prove the fix, and follow the interview’s rules.
Review the first patchPractice interviews
02 Test the boundary, not just the happy path
“What are the boundaries and error conditions?” (opens the source)
A green suite means little if it never tests the contract’s edge.
03 Resolve ambiguity before you optimize
“We want to see your approach to trade-offs like verbosity versus optimization.” (opens the source)
Name the missing requirement before choosing a data structure.
04 Review with evidence a teammate can use
“how well you collaborate with a member of the team.” (opens the source)
Show the failure and a small repair. Keep style comments separate.
05 Choose checks for the actual failure
A retry, a release, or a test needs a specific failure model.
Sources
- Simon Newton, Author, Canva Engineering, at the time. Canva Engineering, “Yes, You Can Use AI in Our Interviews. In fact, we insist”, Jun 11, 2025. Open the source. Checked against the published text.
- Microsoft, Technical interview guidance, at the time. Microsoft Careers, “Technical interviews”, Undated. Open the source. Checked against the published text.
- Atlassian, Engineering interview handbook, at the time. Atlassian Careers, “Atlassian engineering interview handbook”, Undated. Open the source. Checked against the published text.
- GitLab, Technical interview handbook, at the time. GitLab Handbook, “Technical Interviews”, Undated. Open the source. Checked against the published text.
- Addy Osmani, Morgan Westlee Lunt, and Qing Wang, Authors, Claude Code engineering article, at the time. Claude engineering blog, “How the Claude Code team does code review”, Oct 15, 2026. Open the source. Checked against the published text.
Last reviewed Nov 10, 2026