Generative AI Agent Engineer

Outlier needs Generative AI Agent Engineers to train proactive systems using technical feedback, agent debugging, and complex workflow integration remotely.

This freelance opportunity supports leading AI organizations developing large language models into multi-step agents. Contributors evaluate system behavior, improve agent performance, and examine real interactions with tools including Supabase, Gmail, and APIs. The work requires at least two years of relevant experience and supports flexible scheduling, with contributors commonly working five to ten hours weekly, up to forty.

Key Responsibilities:

• Create rubrics that evaluate agent outputs and behaviors
• Investigate traces from multi-turn agent interactions
• Test agents against complex real-world workflows
• Assess coordination, reasoning, and system failures

Requirements:

• Two or more years of backend engineering or AI automation
• Experience integrating complex systems and modular services
• Production software development beyond mocked environments
• Clear, detailed feedback on complex technical behaviors
• Applicants must be authorized in their country

Use agent debugging, systems integration, and precise evaluation skills to help Outlier train advanced generative AI systems.