Remote Data Annotator Jobs Barcelona

Rex.zone is hiring a full-time Remote Data Annotator to produce and validate labeled datasets that improve AI/ML systems across NLP, computer vision, and multimodal workflows. You will follow detailed rubrics, resolve edge cases, and support LLM training pipelines through RLHF, prompt evaluation, QA evaluation, and content safety labeling.

What You Will Do

• Perform data labeling for text, images, audio, and video using project tools
• Create RLHF preference comparisons, rankings, and evaluation examples
• Execute QA evaluation checks for accuracy, consistency, and ambiguity handling; report systematic issues
• Conduct prompt evaluation and response grading for large language model evaluation
• Apply named entity recognition (NER) tagging, classification, and extraction per guidelines
• Complete computer vision annotation (bounding boxes, polygons, keypoints, segmentation) as needed
• Perform content safety labeling for categories like hate, harassment, sexual content, and self-harm
• Maintain audit trails, follow secure data handling procedures, and meet throughput targets

Required Qualifications

• Experience in data annotation/data labeling with measurable quality outcomes
• Ability to follow complex rubrics and apply annotation guidelines consistently
• Familiarity with QA concepts (e.g., inter-annotator agreement, error categorization)
• Comfort working with NLP and/or computer vision datasets
• Strong written communication for issue reporting and adjudication notes
• Ability to work full-time remotely with reliable connectivity and secure work practices

How Quality Is Ensured

• Guidelines compliance checks, gold-standard calibration, peer review, and adjudication
• QA evaluation metrics focused on accuracy, consistency, and reducing error patterns

Pay Competitive hourly rate: $30–$50/hour (based on project scope and performance).

Apply via Rex.zone and highlight relevant labeling, RLHF, QA evaluation, prompt evaluation, NER, computer vision, and content safety experience.