Remote Data Labeling Jobs Paris

About the role

Rex.zone is hiring for remote data labeling jobs aligned with Paris talent, focused on training-data creation for AI/ML systems. You will label text, images, audio, and video to support large language model evaluation, RLHF (Reinforcement Learning from Human Feedback), prompt evaluation, and content safety labeling. You will follow annotation guidelines, document edge cases, and contribute to calibration sessions to improve training data quality and model performance.

Responsibilities

• Perform data annotation across text, image, audio, and video tasks using production labeling tools
• Execute RLHF workflows including pairwise preference ranking, justification tagging, and rubric-based scoring
• Conduct prompt evaluation and response grading for accuracy, instruction-following, and safety
• Apply content safety labeling policies (toxicity, self-harm, hate, sensitive attributes)
• Maintain annotation guidelines compliance; escalate ambiguous cases with clear examples
• Participate in QA evaluation, audits, and disagreement resolution to improve training data quality
• Track labeling throughput and accuracy metrics while meeting full-time delivery goals
• Provide feedback to improve rubrics, taxonomies, and labeling ontology for scalable LLM training pipelines

Requirements

• Experience in data labeling/data annotation operations (text and/or computer vision annotation)
• Familiarity with NLP tasks (e.g., named entity recognition, intent classification, semantic similarity)
• Ability to learn and apply strict rubrics for RLHF and QA evaluation
• High attention to detail and consistency under changing guidelines
• Comfort handling sensitive content in content safety labeling workflows
• Clear written communication for edge-case documentation and reviewer notes
• Reliable internet and availability for full-time remote work

Pay

Hourly range: $30–$50 USD (based on annotation experience, QA performance, and domain fit across NLP, computer vision, RLHF, and content safety).