Job Description & How to Apply Below
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.
To View & Apply for jobs on this site that accept applications from your location or country, tap the button below to make a Search.
(If this job is in fact in your jurisdiction, then you may be using a Proxy or VPN to access this site, and to progress further, you should change your connectivity to another mobile device or PC).
(If this job is in fact in your jurisdiction, then you may be using a Proxy or VPN to access this site, and to progress further, you should change your connectivity to another mobile device or PC).
Search for further Jobs Here:
×