Autonomous Agent Evaluation Engineer

Evaluate autonomous agents remotely for Outlier, applying backend engineering and technical feedback skills to OpenClaw workflows. This is freelance work.

Outlier collaborates with leading AI organizations to train large language models as proactive, multi-step agents. As an Agentic AI Engineer, you will assess complex system behaviors through structured evaluation and debugging. The work includes designing feedback for architectural workflows and identifying subtle failures. Candidates need at least two years of backend engineering, AI automation, or complex systems integration experience.

Key Responsibilities:

• Create evaluation rubrics for agent performance
• Debug traces from autonomous agent interactions
• Stress test agents against complex workflows
• Identify privacy leaks and authority escalation
• Detect indirect prompt injection failures

Requirements:

• Maintain production-grade modular software
• Work with live tools and non-mocked environments
• Handle multi-turn system interactions
• Deliver clear, high-density technical feedback
• Two or more years of relevant experience

Apply careful agent evaluation, debugging, and systems expertise to improve OpenClaw and other advanced generative systems.