Build, test, explain
Clarify an original problem, implement a solution, test boundaries and explain complexity. Choose your language and practice time in the workspace.
Open coding round →BIG TECH / COMPANY-STYLE PRACTICE
Atlassian-style preparation
Practice data structures, code design and changing system requirements. Explain tradeoffs and how you adapt.
14 original exercises selected by ResumeWorlds. These are preparation suggestions, with no claim that Atlassian asked these questions or endorses ResumeWorlds.
Why these topics?
The official handbook describes data structures, code design, system design and manager conversations. These original exercise selections are editorial judgment, not actual Atlassian questions or a guaranteed loop.
Read the public preparation source →Source reviewed 2026-10-11. Role and interview formats can vary; confirm your actual process with your recruiter.
A useful practice round
Choose a coding exercise, explain the approach aloud, implement it and test edge cases. Then explain time and space complexity. For design, write assumptions, compare alternatives and respond to the pressure round. Runtime results require the isolated runner to be available.
Customize and save your round sequence →
Read, repair, verify
Work with intentionally faulty code. Find a counterexample, review your diff and verify the repair with the isolated runner. Record how you evaluated AI coaching.
Open read & repair →Design under pressure
State assumptions, compare architectures and respond to a changing constraint. Existing cited coaching stays separate from executed code correctness.
Open design scenario →Company-style selections are editorial practice suggestions. View topics and source →
14 of 14 original questions
Pair a mentoring budget
Count index pairs i < j whose activity costs add to the target budget. Equal values at different indexes count separately.
Minimum review capacity
Review n positive workloads in order across at most d days. Each day takes a contiguous group. Find the smallest daily capacity that can complete every workload.
Validate nested actions
Positive IDs open actions; negative IDs close the matching positive ID. Actions must close in reverse opening order. Print 1 if the entire sequence is correctly nested and fully closed, otherwise 0.
Fewest connection hops
Given an undirected unweighted graph, return the fewest edges from source to target, or -1 if unreachable. Duplicate edges and self-loops are allowed.
Count target-sum segments
Count nonempty contiguous segments whose sum equals the target.
Front after queue commands
Each command pair is (1,x) to enqueue x or (2,0) to remove the front if present. Return the final front, or -101 for an empty queue.
Strict search-tree check
Print 1 if the supplied binary tree satisfies strict BST ordering, otherwise 0. Every value in a left subtree must be smaller and every value in a right subtree larger; duplicates invalidate it.
Count valid dependency orders
Count vertex permutations in which every directed edge u→v places u before v. Cyclic inputs have zero orders.
Cheapest directed service route
Return the cheapest directed source-to-target route cost, or -1 if unreachable. Source equals target has cost zero.
Critical network edges
Count undirected edges whose individual removal increases the connected-component count. Parallel edges are separate: removing one may leave another route.
Fair API rate limits
Limit API usage across free and paid tenants while retaining clear retry behavior.
Concurrent document editing
Support shared document editing, offline reconnect and version history.
Distributed task leases
Coordinate workers so a task has one valid lease holder at a time.
Online schema migration
Migrate a large user table to a new schema while the application keeps writing.