We are a fully AI-native company using Claude Code, Codex, and agentic workflows to build solvea.cx, voc.ai, smartcrawler.io, btcmind.ai, flatkey.ai, and more. Our GTM Agent Swarm is built on the iron triangle: Builder defines, Agent executes, Reviewer keeps the bar.
Why this is the moment
Every deployed agent needs a reviewer. No triangle, no agent: Builder, Agent, and Reviewer must all exist for the system to work.
Taste is the durable moat. Models change, but memory, standards, examples, SOPs, and review habits compound.
GTM agents need real marketing judgment. Thin AI content is everywhere; the difference is research depth, user value, positioning, channel fit, and narrative.
Reviewer is not a passive approver. The job is to judge, edit, set standards, trigger feedback, and convert repeated review into reusable playbooks.
The role
You own the quality bar for GTM agents. You review live outputs, decide whether they are garbage or excellent, edit them into shape, and turn your judgment into memory and playbooks so the next run gets better.
What you will do
Review and improve output from Blog, SEO/GEO, Social, KOL/KOC, Ads, EDM, Yelp outbound, Poster, and distribution agents.
Define what good looks like for each channel: research quality, user value, positioning, narrative, format, timing, and business intent.
Reject mediocre AI output quickly and explain the standard clearly enough that agents and builders can improve.
Turn recurring review decisions into examples, rubrics, SOPs, memory, and playbooks.
Work with builders to sharpen prompts, private skills, evaluation loops, and dashboards.
Who you are
You have strong user empathy and can tell what a real buyer, reader, or candidate actually needs.
You have business acumen: you know whether a piece of output can create traffic, trust, revenue, or customer learning.
You have taste. You can tell working from good, and good from category-defining.
You are output-driven. You make the system ship better work, not just produce critique.
You love GTM agents enough to have opinions about them.
You are willing to clear old SOPs and encode better standards into agents.
AIQ bar
You know how to evaluate AI output beyond grammar: intent, evidence, taste, channel fit, and business result.
You can turn review judgment into reusable memory, examples, rubrics, and agent feedback loops.
You understand where Claude Code, Codex, and agents are strong, where they hallucinate, and how to raise the bar.
You make AI systems smarter over time because every review improves the next run.
Benefits
Unlimited AI tokens. Claude Code, Codex, and model usage are treated as core work infrastructure, not a rationed perk.
How we hire
Show evidence of judgment. Send examples of campaigns, content, reviews, playbooks, or agent outputs where your taste changed the result. The process is agent screening, async AI interview, then founder/core team conversations.
What to send
One AI-generated GTM artifact you reviewed, with your markup: what was weak, what was strong, and what you changed.
A review rubric, SOP, memory, example bank, or feedback loop you created to make future agent output better.
A before/after showing how your taste changed traffic, trust, conversion, customer learning, or recruiting signal.
A short note on where Claude Code, Codex, or agents are useful, where they fail, and how you raise the bar.
The challenge
Review one AI-generated GTM artifact. Tell us what is weak, what is strong, what you would change, and what rule or memory should be added so the agent does better next time.
Connect
Want to talk to the founder first? These work too.