Remote RN: AI Training & Model Evaluation Expert

Prolific Academic Ltd; United States; Remote; USD 74,000 - 105,000; Full time

Job summary

Prolific is seeking Registered Nurses to help train and evaluate AI models. A quick test screens for suitability; successful participants join as Domain Experts and get paid per AI task.

Tasks typically require one hour of uninterrupted work, with many being shorter, and the position offers flexibility and work-from-home options. Experts review AI-generated clinical scenarios, rate them on accuracy, safety, and reasoning, and may compare model answers to justify the best response.

Qualifications

• Verified status as a registered nurse with current license/registration in good standing.
• Recent clinical experience and ability to evaluate clinical reasoning.
• Willingness to complete a short skills/eligibility screener to join the Domain Expert pool.
• Strong attention to detail and ability to focus on complex tasks for up to one hour.
• Reliable, fast internet connection and access to a computer.
• Willingness to self-declare earnings as a self-employed participant.
• PayPal account required to receive payments.

Responsibilities

• Review AI-generated responses to clinical scenarios and rate them for accuracy, clinical appropriateness, safety, and reasoning quality.
• Compare multiple model answers and select/justify the best response.
• Write improved exemplars, rationales, or structured feedback to help models learn where they fall short.

Skills

Registered nurse

Clinical experience

Attention to detail

Internet access

Self-employment

Job description

Prolific is seeking Registered Nurses to help train and evaluate AI models. A quick test screens for suitability; successful participants join as Domain Experts and get paid per AI task.

Tasks typically require one hour of uninterrupted work, with many being shorter, and the position offers flexibility and work-from-home options. Experts review AI-generated clinical scenarios, rate them on accuracy, safety, and reasoning, and may compare model answers to justify the best response.