Process
How hiring here works
Five stages, no bidding, and a shortlist you can interrogate rather than take on trust.
You send a brief, our matching engine returns a ranked shortlist with a visible score breakdown, you review each candidate vetting record, then interview and run a short paid trial before committing. A shortlist typically arrives within two business days of a complete brief.
The five stages
Send a brief
Describe the work in plain language: what you are building, the skills involved, your budget ceiling, the time zone you need covered and when you want to start. Free text is fine — our intake agent parses it into a structured query and shows you what it extracted, so you can correct it.
Receive a scored shortlist
Our matching engine scores every available engineer against the brief on five weighted components: skill overlap at forty percent, role fit at twenty, seniority fit at fifteen, rate fit at fifteen and time-zone overlap at ten. You see the breakdown per candidate, not just a ranking.
Review the vetting evidence
Every profile carries the record behind it: which work samples were reviewed and what was found, the structured interview scorecard, and verified reference notes. If a candidate has a gap relevant to your brief, it is shown rather than omitted.
Interview and run a paid trial
Interview the shortlist yourself, then start with a short paid engagement on real work — typically one to two weeks. A paid trial is the only assessment that reliably predicts the following six months, and it is fair to the engineer as well.
Onboard and scale
Contracting, invoicing and time-zone arrangements are handled. As the work grows, additional engineers can be matched against the same brief context, so the second hire does not start from a blank page.
Why the scoring is visible
Most platforms present a shortlist and ask you to trust it. That is a reasonable request only if you can see what produced it. Our matching is deterministic — the same brief always yields the same result — and every candidate arrives with the five component scores that placed them there.
This matters practically, not just philosophically. If a candidate ranks highly on skill overlap but poorly on time-zone overlap, that is a decision for you to make with the information in front of you, not one an opaque algorithm should quietly make on your behalf.
Where the language model is, and is not
An agent built on the OpenAI Agents SDK reads your brief and converts it into a structured query. That is the only place a model is involved. The ranking itself is deterministic code with published weights, because a shortlist that changes between identical requests is not a shortlist, it is a suggestion.
Frequently asked questions
How long does hiring take end to end?
A shortlist usually arrives within two business days of a complete brief. Interviews and a paid trial typically add one to two weeks. Most engagements start between one and three weeks after the initial brief, and the limiting factor is almost always how quickly your side can interview.
What makes a good brief?
Specificity about the problem rather than the person. "We need someone to fix an agent system that loops on tool failures" produces a far better shortlist than "we need a senior AI engineer". Include your budget ceiling and required time-zone overlap, since both are scored directly.
How does the matching engine work?
It scores every available engineer on five weighted components — skill overlap, role fit, seniority fit, rate fit and time-zone overlap — and ranks by total score. It is deterministic, so the same brief always produces the same shortlist. A language model parses your brief, but it never does the ranking.
Do I have to run a paid trial?
No, but we recommend it strongly. Interviews measure how someone talks about work; a short paid engagement measures how they do it. It also gives the engineer a fair view of your codebase before committing to months of it.
What happens if the engineer is not right?
Tell us during or immediately after the trial and we re-match at no additional platform fee. We treat a bad match as a failure of our process, which is precisely why the vetting evidence is shown to you upfront rather than summarised.