Job Description
**Software Engineer**
Remote (US or UK) · Part-time contract · 10 hrs/week · $3,000 $4,000/month fixed rate (USD) · Flexible schedule
**About the role**
Frontier AI models can now write an entire feature in minutes. Someone has to check whether that code actually works, and that someone is you.
You'll run agentic coding sessions on real engineering tasks, review what the models produce, and push them until they fail in interesting ways. Every failure you document helps train the next generation of coding agents.
This is a part-time role, about 10 hours a week on your own schedule, which fits alongside a full-time job or other contracts.
**What you'll do**
- Vibe-code with frontier models on real tasks, like fixing a bug, extending an API, or shipping a feature, and see how far the model gets on its own.
- Review the AI's work against production standards: find the bugs, explain why they happened, and rate the quality of the model's reasoning.
- Red-team the models to expose unsafe code, faked test passes, and confidently wrong solutions before real users run into them.
**Who does well here**
Software Engineers, Backend Engineers, Full-Stack Developers, and DevOps Engineers who enjoy taking code apart as much as writing it.
**What we look for**
- Professional experience shipping production code, or substantial open-source contributions.
- Fluency in at least one major stack. Most tasks use Python, C++, Rust, Go, or JavaScript.
- Clear written English. Your explanations are the training signal.
- Based in the US or UK with the right to work there.
- No AI or machine learning experience needed. If you can spot bad code, you can spot bad AI-generated code.
- No degree required. We care about what you can do, not where you learned it.
**Compensation**
$3,000 $4,000/month fixed rate, paid in USD, for approximately 10 hours per week, depending on experience.
We review every application.