Skip to the content.

Interview sprint

210 minutes across 8 steps. The shortest path to being dangerous in a loop: one question per round type, chosen because each one is where candidates most often come apart.

Interview sprint 0 of 0 done (0%)
  1. Where does an atom stop? quiz · easy · 8 min

    The component-boundary question, asked four ways.

  2. Memoize a derived selector coding · medium · 25 min

    A JavaScript round that is really about when caching pays.

  3. Build a loading Button atom ui-coding · easy · 30 min

    A UI round in miniature.

  4. Build an accessible combobox ui-coding · hard · 45 min

    A UI round with the keyboard requirement people skip.

  5. Pick a rendering strategy quiz · medium · 10 min

    The first decision in any design answer.

  6. Design an embeddable widget system-design · hard · 45 min

    A full design round with real constraints.

  7. Where does the session actually live? quiz · hard · 12 min

    Where the plausible answer is the insecure one.

  8. Normalise a nested API response coding · hard · 35 min

    The data-shape question behind half of 'why is this slow'.

Three and a half hours, eight questions, one per thing a loop actually tests. This is not a substitute for the longer plans — it is what you do the weekend before, or what you do first to find out which of the longer plans you need.

Each question is here because it is a common failure point rather than a common question. The combobox catches people who have never built for a keyboard. The session quiz catches people whose instinct is the convenient answer. The normalisation exercise catches the data-shape decision hiding behind half the “why is this slow” conversations.

Do them under time and out loud. If you find yourself going quiet on the hard part, that is the finding — talking while stuck is the skill being graded, and it is trainable in about a week.