How it works
One connected path from the invitation to the hiring decision. Candidates get a realistic task with clear rules. Reviewers get the code, tests, and reasoning behind it. Your team makes the call.
01
Pick a format and a task for the role, state the AI rules, and send the invitation.
02
Realistic engineering work in the browser, with the checks they need to verify it.
03
Code, tests, prompts, replay, and the candidate’s own explanation in one packet.
04
Record the decision, or bring the thin parts into a short live interview first.
What candidates experience
The process respects the candidate’s time and shows them exactly what is expected, so the evidence means the same thing for everyone.
Candidates open a private link. No account to create. The invitation says what the role is, how long the task takes, whether AI is allowed, and what is recorded.
They choose when to start inside the window your team set, so the work fits their schedule instead of yours.
Instead of a puzzle, candidates get the kind of work the job involves: a small service with tests, a bug to find and fix, a change to review, or an AI proposal to challenge.
The brief, the starter code, and the checks sit side by side in the browser. Work autosaves, and a dropped connection resumes where they left off.
When your rules allow it, candidates use AI in the open. They run the checks, read the output, revise, and run them again.
Candidates submit before the time runs out, with a short note on what they changed, what they verified, and what they would still want to check. If time ends first, the work saved so far is submitted for them.
They receive a confirmation by email, so nobody is left wondering whether the submission went through.
After a reviewer approves it, your team can release a private Candidate Growth Report: evidence-backed strengths and useful next steps, without exposing your internal scoring.
Want to see the task itself? Explore the candidate journey.
What reviewers receive
A single score hides how the work was done. Reviewers see the reasoning, not just the outcome, and the hiring decision stays with people.
Each submission arrives with the code, the test results, the prompts and AI suggestions, a replay of the session, the candidate’s decision note, and integrity context such as focus and paste events.
Integrity events are context for a person to read, never a cheating verdict on their own.
Executable checks show what the work proved. Where the task could not prove something, the packet says so, so a strong local result is never mistaken for a guarantee.
Every gap becomes a precise question for a short live interview. Interviewers arrive knowing what to ask instead of repeating the whole assessment, and candidates get the chance to show depth where the evidence was thin.
Reviewers add notes, complete the scorecard, and record the decision. The assessment, the evidence, the live interview, and the outcome stay connected in one trail your team can share or keep internal.
Want to read a real packet? Open the sample report.
Three formats
Each format is an independent option with no required sequence. Use one, or combine them for the roles that need it.
Candidates complete a focused task on their own schedule. Your team reviews the evidence when it suits them.
ExploreAn engineer from your team explores decisions with the candidate in a shared workspace.
ExploreProposed paid sessions for finalists alongside potential teammates, with pay regardless of the hiring decision.
ExploreSee it with your role
A 15-minute walkthrough with your open role: the task, the evidence packet, and the reviewer workflow your team would use.