Preparing your design workspace

Loading the problem and your latest saved attempt…

Design an Employee Treasure Hunt App

Help employees meet in small teams, pool clues on a shared board and solve a treasure hunt together.

Problem statement

Design an application that helps employees meet one another. For this exercise, use a cooperative treasure hunt as the MVP: colleagues form a team, combine clues and solve the final puzzle. Include the user flow, backend data, API contracts, storage and a scaling plan.

For this exercise, assume employees already have stable company-directory identities. An organizer drafts a hunt with one to ten required clues, valid clue tokens and one predetermined answer. Publishing seals that definition and opens the hunt; rules cannot change underneath its teams.

An organizer can close an open hunt. Puzzle generation, chat and physical-location checks are outside scope.

For this exercise, assume an employee belongs to at most one team in a hunt. A forming team has at most four members, and membership can change while it is forming. Any team member can start it with two to four members; start freezes that membership.

Members of an active team contribute valid clue tokens to a shared board. Each required clue counts once per team. A correct final answer completes the team only after all required clues have been recorded, and a team completes once.

For this exercise, assume current server state governs every new action. A closed hunt accepts no genuinely new action. Replaying a recorded successful action with the same identity and input returns its original committed result, including after closure; changing the input under that recorded identity is a conflict.

Board views may lag within the stated target, but a stale screen must not authorize a write. Design a lobby, a shared clue board showing teammates/progress, and a stable completion view.

What to build

  1. Let an organizer draft, publish/open and close a hunt while keeping a published definition fixed.
  2. Let employees form teams and start with two to four members, enforcing at most one team per employee and freezing membership at start.
  3. Let active teammates record valid clues once each and submit the known final answer after collecting every required clue, with one completion per team.
  4. Show a lobby, shared teammate/clue progress and a stable completion result, including recovery of a recorded successful action after a lost response.

At scale

  • For this exercise, assume a company directory of 20000 employees and at most 2000 participants in one hunt.
  • For this exercise, assume peak traffic of 500 board reads and 100 state-changing commands per second.
  • For this exercise, show committed clue progress on teammates’ board views within five seconds. Capacity, membership, completion and closure checks must use current authoritative state.
  • For this exercise, a hunt definition has one to ten required clues and a team has at most four members. Retain the committed results needed to replay recorded successful actions for the life of the activity record.