Evaluate autonomous agents remotely using backend engineering, AI automation, and complex systems integration skills. Outlier offers freelance work with flexible scheduling.
This prospect supports Outlier projects that train large language models to function as proactive, multi-step agents. Remote contributors design feedback around real-world architectural workflows and system behavior. The work suits software developers and orchestration specialists with at least two years of experience. Applicants are accepted from Australia, Canada, Finland, France, Italy, Norway, the United Kingdom, and the United States.
Key Responsibilities:
- Stress test autonomous agents across real-world workflows
- Debug traces from multi-step agent interactions
- Write rubrics evaluating agent performance
- Identify failures in complex system behavior
- Help improve generative systems through human feedback
Requirements:
- At least two years of relevant professional experience
- Experience building and maintaining production-grade software
- Comfort with live tools and non-mocked environments
- Understanding of modular services and system integration
- Attention to detail when reviewing technical behavior
Use autonomous-agent evaluation, backend development, and system-integration skills to support Outlier’s remote AI training projects.