Remote AI Trainer & Evaluator - Flexible, Task-Based

prolificacademicltd; United States; Remote; USD 21,000 - 34,000; Full time

Benefits offered by this job

Flexible hours

Remote work

Task-based pay up to $25/hr

Job summary

Prolific Academic Ltd in the United States is seeking experienced AI trainers to evaluate and improve cutting-edge models. You join a self-employed contributor pool to complete focused paid tasks from anywhere, with remote, flexible hours and task-based pay.

Responsibilities include analyzing, editing, and annotating model outputs, evaluating prompts, flagging harmful content, and contributing data to benchmark and improve model quality. A PayPal account is required for payments.

Qualifications

• Verifiable professional experience as an AI trainer or in a closely related annotation, evaluation, or data-quality role.
• Willingness to complete a 10–15 minute skills assessment to qualify for the participant pool.
• Strong attention to detail and the ability to stay focused on demanding tasks for up to one hour at a stretch.
• A dependable internet connection and a personal computer suitable for online work.
• A PayPal account to receive per-task payments.
• Comfort operating as a self-employed contractor and self-reporting income.

Responsibilities

• Carry out AI training tasks such as analyzing, editing, and writing annotations on model outputs
• Judge the performance of AI systems across a range of prompts and provide constructive evaluations
• Flag harmful, biased, or inaccurate outputs to support safer model alignment
• Contribute data that helps researchers benchmark and improve model quality

Skills

Attention to detail

Remote work discipline

Independent contractor

Job description

Prolific Academic Ltd in the United States is seeking experienced AI trainers to evaluate and improve cutting-edge models. You join a self-employed contributor pool to complete focused paid tasks from anywhere, with remote, flexible hours and task-based pay.

Responsibilities include analyzing, editing, and annotating model outputs, evaluating prompts, flagging harmful content, and contributing data to benchmark and improve model quality. A PayPal account is required for payments.