An agent that operates makes decisions. Someone has to think rigorously about what it should decide alone, what is never allowed and how it explains what it did. That is the job.
The role
- Write the agents' judgment: what they decide, what they escalate and what is never allowed.
- Design tests that check whether an agent reasoned well, not only whether it got it right.
- Review how an agent explains its decisions and whether that is enough to sign them.
- Bring that judgment into the governance our clients read.
Your first 90 days
- Day 30
You learn the agents in production and their governance; you review their decisions.
- Day 60
You write the judgment for a complete agent with its edge cases.
- Day 90
Your reasoning tests run on the agents before every change.
Stack and tools
Language modelsAgent logsGovernance documents
What we look for
Must have
- A philosophy degree, finished or nearly finished.
- You write precisely and without filler.
- You care about ethics applied to systems that act, not in the abstract.
- You argue with reasons and change your mind with reasons.
Nice to have
- Logic, philosophy of mind or applied ethics.
- Curiosity about how language models work.
- Some programming.
What working here is like
What you build goes into products that operate real organizations. You work directly with the founders, with no layers between your code and production, under the same rule we give the intelligence: a person signs what is critical.