Remote Generative AI & RLHF jobs

What this work can involve

Generative AI and reinforcement learning from human feedback work uses people’s judgments to help prepare and evaluate AI systems. Depending on the individual role, you may write prompts, compare and rank model responses, review generated content, identify weaknesses through red-teaming, or assess whether outputs meet the stated task. The work can also involve documenting judgments or applying defined criteria to model responses. Read each live posting carefully because responsibilities and requirements are specific to that role.

• Write prompts designed to test or elicit particular model behavior.
• Rank, compare, or review responses for quality, relevance, accuracy, or adherence to instructions when a posting calls for that work.
• Use red-teaming to probe weaknesses and surface problematic model outputs.

How hiring works through OpenTrain

OpenTrain hires for the public roles available through its site. Candidates apply through OpenTrain, and OpenTrain may place or refer successful candidates to the client whose project they will support. The individual posting is the source of truth for the role’s eligibility, responsibilities, qualifications, experience, pay, hours, schedule, location rules, and any client-specific requirements.

• Review the complete live posting before deciding whether the role matches your situation.
• Apply through OpenTrain for roles that fit your interests and the requirements stated in the posting.