Skip to content

Roles

We assess the people who build the product

Six roles, one rubric. We only cover work where the job is deciding under ambiguity and the brief always arrives wrong — so backend, data and infrastructure are deliberately out of scope.

That limit is the point. A library of four hundred shallow tests is a different product, and it already exists.

Request access to the beta

The same instrument, six times

Every role gets 12 written situations in about 15 minutes, across 4 short parts. What changes between them is the situations themselves — a product manager is asked about a deal lost to a missing feature, a UX writer about an error message that is covering for a broken flow. The rubric, the scoring and the report are identical.

That is what makes two candidates for the same job comparable, and it is the reason we do not add roles quickly.

Why these six and not four hundred

A role belongs here when the person builds the product experience and the job is deciding under ambiguity. Backend, data and infrastructure are out on purpose: their hard part is verifiable in other ways, and a wide catalogue would make us one more test library.

You pick which capabilities can veto a hire when you set the role up, so the same answers read differently for two different jobs. How the scoring works, and what it costs.

Before you pick one

No, and it is deliberate rather than a roadmap item. Each role has its own bank of situations written for the decisions that job actually makes — a product manager is asked about a deal lost to a missing feature, a UX writer about an error message covering for a broken flow. Running someone through another role's situations would produce a number, and the number would not mean anything.

Pick by what the person will decide, not by what the title says. If they own what gets built and why, that is the product manager assessment even if you call it something else. If they live between design and engineering deciding when a component should exist, that is design engineer. Titles vary by company; the decisions do not.

Only where the job is deciding under ambiguity with a brief that arrives wrong — that is the limit, not a stage we are passing through. Backend, data and infrastructure stay out because their hard part can be verified other ways. If a role you hire for belongs inside that limit and is not here, tell us: the pages for roles we have not built yet carry a direct line for exactly that.

You have an open role right now.

Invite one candidate and read their report the moment they finish.