Assessment that survives being questioned.
A hiring decision, a promotion, a professional credential — each is a decision somebody may challenge. The same controls that let a university defend a result let an employer defend one.
Skills, aptitude and technical evaluation
Twelve question types
Multiple choice and multiple response, true/false, matching, ordering, fill-in, numeric, short and long answer, and file upload. Every type is marked or routed to a human, per question.
Cohorts and windows
Schedule a paper for a group, with a window that opens and closes, per-section timing, and randomisation so two candidates do not see the same order.
Item statistics
p-value and discrimination per question, recomputed from released marks — so a question that is not separating candidates can be found and retired.
What happens to a coding answer, precisely
Coding and SQL questions can be set, sat and marked here. They are not executed. An answer is captured as the candidate wrote it and goes to a human evaluator, with the question, the expected approach and the marking scheme beside it.
We say so plainly because the alternative is a buyer discovering it during a pilot. There is no sandbox in this platform: nothing compiles, nothing runs against test cases, and no score is produced by a machine for a coding answer. If automatic execution is what you need, this is the wrong tool today, and we would rather tell you now.
What you do get is a marked answer with a name against it, a moderation path, and the same two-person release gate as everything else.
Issue a credential, and let anybody check it
A certificate is generated from a released result, carries a code with forty bits of randomness, and has a public address anybody can open without an account.
- Certificate templates with your own signatories
- Batch generation from a released cohort
- Revocation that the public page reports honestly
- A verification portal that needs no login and no data from the checker
Placement and recruiter reporting is planned and not built — nothing in the platform records a placement yet, and the recruitment report says so rather than drawing a chart from an empty table.
See it run on your examination.
Bring a real paper, a real cohort and the constraint you think will break it. Thirty minutes, no slides.
No commitment, and no sales sequence afterwards unless you ask for one.