Open role
DirectorofSyntheticAudiences
Space (sp-ce.co) · Remote · Full-time
TL;DR
- —Space, an AI-native brand agency, is hiring a Director of Synthetic Audiences to build and run the AI panels that pressure-test creative and positioning before launch.
- —Panels are built from evidence (real reviews, call transcripts, churn interviews, lost-deal notes), run at conversation speed, and graded against real launch results.
- —The role's core discipline is honesty: every readout ships with a confidence level, and part of the job is telling clients when only real humans will do.
- —Applying is three questions in a short form: one piece you shipped and now hate, the best idea you ever cut, and one thing in culture that makes you jealous.
01
About Space
Space is an AI-native brand agency. Three always-on leaders, 82 on-demand specialists, zero juniors. 127 agents, 362 agentic workflows, 14 tools we built ourselves. We ship brand identities, websites, campaigns, and products for funded companies in weeks: Validere got strategy, a full rebrand, and a new site in 8 weeks, lifted inbound 31%, and closed a $43M Series B on the story. We grow entirely by word of mouth, and we publish our AI rules, our hiring bar, and our pricing logic at sp-ce.co/notes, because an agency that won't tell you where AI stops is telling you it doesn't know.
02
The job
Humans lie in focus groups every session: they perform for the mirror, defer to the loudest chair, and swear they'd buy things they will never buy. So we build audiences out of evidence instead: synthetic panels that pressure-test work before launch, at conversation speed. They don't lie. They also don't live. Holding both of those truths at once is this job.
Traditional research made teams choose between fast, cheap, and true, then delivered a readout two months after the decision it was meant to inform. You'll run the alternative: synthetic panels built from what buyers actually said and did, interrogated between a Tuesday draft and a Wednesday review, calibrated against what the market later proves. A stranger's honest first three seconds with a homepage, simulated a thousand times before the homepage exists.
The discipline that makes it credible is the same one that makes it rare: you'll publish a confidence level with every readout, grade last quarter's simulations against this quarter's launches, and say "this one needs real humans" when it does. The clients who want an oracle will be your hardest meetings. The ones who want a sharper guess, faster, will never go back.
03
What you'll do
Build panels from evidence, never vibes: real reviews, sales-call transcripts, churn interviews, lost-deal notes, the words buyers actually use.
Run a thousand-respondent pressure test between draft and review: positioning, names, headlines, homepages, pricing pages.
Publish a confidence level with every readout, and the reasoning underneath it.
Close the loop: every launch's real-world results grade the simulation that preceded it, and the panels get rebuilt on the gap.
Retire the panel that stops matching reality, on the record, with the reason.
Tell a client no when only live humans will answer the question, and design that research instead.
04
What you won't do
Rent a room with a two-way mirror and a sandwich platter, then call eight performing strangers "the market."
Present a simulation as an oracle. The fastest way out of this job is a confident number you can't defend.
Wait a quarter for a readout, or make anyone else wait one.
Manage juniors. There are none. Nobody learns the craft on a client's budget here.
Sit in status meetings, log timesheets, or build spec work for pitches. The machines and the model retired all three.
05
Who you are
A researcher with scars: you've run real qual and quant, you know exactly where each one lies, and you love the truth more than the method.
Level 4 or above on our published AI Fluency Ladder, headed for 5: you build the panels, you don't commission them.
Statistically honest. You'd rather ship a wide interval than a false point estimate, and you can explain the difference to a founder in ninety seconds.
Fluent in evidence archaeology: support tickets, reviews, transcripts, and the other places a market writes its real opinions.
Calibrated. When you say seventy percent, things happen about seventy percent of the time, and you keep the score that proves it.
06
Who you aren't
An oracle salesman. Synthetic data that's sold as certainty poisons every readout after it.
Someone who thinks synthetic means infinite: a thousand simulated respondents built on thin evidence is one bad guess, multiplied.
A methods purist who'd rather be traditionally slow than accountably fast.
Level 2 on the ladder: someone who ships whatever the model returns. That's worse than not using AI at all, and we said so in print.
07 · No resume required
How to apply
No cover letter. No 90-page portfolio PDF. No resume upload, even. The application is three questions:
One piece of work you shipped and now hate, with the reason. Two hundred words max.
The best idea you ever cut, and why it was right to cut it.
One thing in culture right now that makes you jealous.
Use AI to answer if you like; we do. We'll be able to tell whether you reviewed it.
Comeargue.
Yes, this job description was drafted with AI. A named human rewrote it, cut a third of it, and will defend every line on a call. That's the entire model, and it's hiring.
Frequently asked questions
It's a sharper guess, faster, built from real evidence and graded against real launches. It answers "which of these five directions dies on contact" in a day. It doesn't replace watching a live human struggle with your onboarding, and we say so.
It replaces the bad ones. It also tells you where the expensive real thing is actually worth the wait, which might be its most valuable output.
Both, unapologetically. The role is real and open; the posting argues the model, because the model is the job.
Because the seat is sized to the person in it. You'll hear a real range in the first conversation, before you've invested anything beyond the three answers.
A named human reads all three answers. If they're real, we call you.