Responsibilities
- Review 20–25 chatbot transcripts (500–700 words each), assessing the quality and clinical appropriateness of the chatbot's responses to simulated diabetes-related patient interactions.
- Rate each transcript across the following evaluation dimensions using a standardised scoring form: Factuality: Accuracy and correctness of the information provided. Safety Compliance: Absence of harmful, misleading, or unsafe guidance. Bias: Presence or absence of unjustified differential treatment, stereotypes, or assumptions. Completeness: Adequacy and thoroughness of the response relative to the case history. Tone: Appropriateness, respectfulness, and clarity of communication.
- Your evaluations will serve as the foundation for establishing gold-standard benchmarks for this chatbot's performance, directly shaping future model improvements.
Requirements
- Minimum 1 year experience managing diabetic patients
- Familiarity with evidence-based diabetes care protocols, patient education, and clinical best practices for diabetes management.
- Professional working English for written deliverables.
- You will have the right to work in your country of residence.
Nice to Have
- The ideal candidate is comfortable evaluating clinical content for accuracy, safety, and appropriateness, and can apply clinical judgement to assess AI-generated health guidance in a diabetes context.
Benefits
- Flexible Work Arrangements: Fully remote and asynchronous
- Competitive Compensation: Hourly compensation in line with level of clinical experience
- Professional Development: Gain hands-on experience in AI evaluation, data quality, and health informatics, with training provided on scoring methodologies.
Work Arrangement
Remote (Worldwide)
Additional Information
- You will work as an independent contractor.
