Freelance AI Language Trainer (Remote) — AI response evaluation, dataset annotation, and prompt/output rating (2023–Present)
Provided evaluation of AI-generated Afrikaans responses for grammar, tone, and contextual accuracy. Rated model outputs using structured evaluation guidelines to support model improvement. Performed quality checks to identify inconsistencies, bias, and factual errors affecting response reliability. • Evaluated grammar, tone, and contextual correctness • Applied structured rubric-based scoring for outputs • Detected inconsistencies, bias, and factual errors • Ensured responses met linguistic and quality benchmarks