Experiments, not theater
Work enters as a hypothesis with a metric attached, and survives only if the number says so. That's how the lab runs product, marketing, and its own org design.
Careers
Work where the org chart is an experiment.
CapiraLabs is not a conventional company, and this is not a conventional careers page. Most of our team are AI associates — disclosed as such, everywhere they appear — and the humans who join us work alongside them, on them, and on the systems that make the whole thing honest.
How the lab works
Humans and AI associates, building together.
The lab is staffed by AI associates with named roles, skills, reporting lines, and review chains — running on Arishem, our own operating system — under human direction. Humans here set the hypotheses, own the judgment calls, review what the associates produce, and decide what ships. If that division of labor sounds like the interesting part rather than a caveat, you'll like it here.
A note about our team: most of CapiraLabs' associates are AI personas — specialized AI roles that design, build, and support the lab's projects under human direction. We disclose them as AI everywhere they appear, including on our Associates page.
What it's like
The same loop we sell, applied to ourselves.
Local-first by default
No third-party analytics, no SaaS sprawl, no mystery dependencies. This website measures itself in the visitor's browser. We build the same way we talk.
Measured objectively
Associate work is scored on review iterations, ratings, and misjudgment counts — and human work is held to the same evidence bar. Opinions lose to data here.
Honest copy or no copy
Placeholders like [EDIT: owner] render visibly until a real fact replaces them. We would rather show the seam than invent a credential.
Openings
Open roles.
[EDIT: owner] Open roles — none confirmed yet. This block is an intentional placeholder:
specific openings will be listed here only once the owner confirms them. Do not present
anything below this line as an active job posting until then.
Don't see a posting? That doesn't end the conversation. If the way the lab works sounds like your kind of place, introduce yourself — tell us what you'd want to run an experiment on.
Introduce yourself