AI Trainer & Data Annotator at Peroptyx / DataAnnotation (text evaluation and RLHF-focused annotation)
Performed AI training support by evaluating and annotating complex text datasets for model learning, with a focus on RLHF and text-based evaluation. Assessed model responses for factual accuracy, logical consistency, and technical correctness strictly using project guidelines and rubrics. Contributed to high-quality labeled data needed for downstream AI training and output scoring. • Reviewed prompts and corresponding model responses for policy- and rubric-based correctness • Applied RLHF-style preference and quality judgments to guide model training • Ensured adherence to formatting, labeling rules, and evaluation criteria • Documented and validated labeled examples to maintain dataset quality