Evaluate autonomous AI systems from a flexible remote setting using technical feedback, agent debugging, and production-grade software expertise.

Outlier collaborates with leading AI organizations to improve agents like OpenClaw through human feedback. This freelance work supports large language model training for proactive, multi-step systems. The focus includes architectural workflows, live interactions, and complex system behavior.

Key Responsibilities:

- Develop rubrics for evaluating agent outputs
- Debug traces from multi-step agent interactions
- Stress test autonomous agents for subtle failures
- Assess system behavior with explicit technical feedback

Requirements:

- Two or more years in backend engineering
- Experience with AI automation or systems integration
- Maintain modular, production-grade software
- Build for live, non-mocked environments
- Identify privacy leaks and prompt injections

Use systems integration and agent evaluation skills to help improve OpenClaw and other advanced generative systems.

📌 Autonomous Agent Evaluation Engineer (London)
🏢 Outlier
📍 London