AI Coding Interviews

The coding interview is changing.

AI now writes the code that coding interviews used to test. What’s left to test is how you work with it.

Start practisingFrom $19 a month
The practice workspace: files, editor, terminal with tests passing, and the running app
Your repository, Claude Code in the terminal, the running app and the product manager.

What gets tested now

  • Finding out what matters

    The brief is never the whole story. Ask the questions that change the design before you build.

  • Directing the AI

    Give Claude Code a plan and the constraints, not a wish.

  • Checking its work

    Run it, test it, and catch what it got wrong.

  • Choosing what to build

    Read the evidence, weigh what points the other way, and pick the change worth making.

So practise exactly that.

Each round is a real repository and Claude Code in your own sandbox, with a product manager who only answers what you ask. Fix a bug, build a feature, or take a product challenge and decide what’s worth building.

Then a report on how you worked

Every skill is scored from your own session: the questions you asked, the prompts you gave, the checks you ran. Each score quotes the moments it’s based on and tells you what would have earned the next level.

Late fees that look too high96.6 / 100
  • Requirement discovery3 / 4
  • Verification4 / 4
  • Outcome4 / 4

Next time: also ask whether some members or items are exempt, and whether fees already charged need correcting. Then say back what you will and won’t change.

An example report from one of our own practice rounds.

From $19 a month.

No free tier: every round runs a private sandbox and real Claude Code. Every question is original, never taken from a company’s interview. Cancel any time.

Start practising