Build and evaluate autonomous agents remotely with OpenClaw, using backend engineering and AI automation expertise. Freelance schedules remain versatile.
Outlier seeks Agent Builders to help train large language models as proactive, multi-step agents. You will shape architectural workflows through agent evaluation, debugging, and stress testing. The work suits software developers and orchestration specialists with production-grade systems experience.
Key Responsibilities: Write evaluation rubrics for autonomous agent behaviors
Debug agent traces across multi-turn interactions
Stress test agents in complex workflows
Provide technical feedback on system behaviors
Requirements: Two or more years in backend engineering, AI automation, or systems integration
Build production-grade software with modular service separation
Work with live, non-mocked environments and multi-turn systems
Analyze privacy leaks, authority escalation, and prompt injections
Applicants accepted from Australia, Canada, Finland, France, Italy, Norway, United Kingdom, or United States
Apply OpenClaw architecture, agent debugging, and technical evaluation skills to Outlier’s autonomous AI projects.