Remote Data Labeling Specialist (Brussels)

About The Role

Remote data labeling professionals in Brussels support Rex.zone AI training workflows by creating and validating high-quality labeled datasets for NLP, computer vision, and LLM training pipelines. This full-time role focuses on annotation guidelines compliance, training data quality, RLHF and prompt evaluation, and QA evaluation.

What You Will Do

• Execute data labeling for text, images, and multimodal datasets using tool-based workflows and taxonomy rules
• Perform named entity recognition (NER) tagging and other NLP annotation tasks
• Conduct LLM prompt evaluation and RLHF preference ranking; document clear rationales
• Run QA evaluation on labeled batches to improve inter-annotator agreement and reduce noise
• Apply content safety labeling using policy-driven risk categories and escalation rules
• Track annotation guidelines compliance, resolve ambiguity, and document edge cases
• Report dataset issues and propose rubric or guideline refinements to stakeholders
  
  

Required Qualifications

• Mid-Senior experience in data labeling, data annotation, QA evaluation, or dataset operations
• Proven ability to maintain training data quality under throughput targets
• Familiarity with NLP/LLM evaluation concepts (prompt evaluation, RLHF, rubric-based grading)
• Strong written reasoning for documenting disagreements and labeling decisions
  
  

Tools & collaboration

Work in structured labeling queues with audit trails, batch reviews, and measurement of agreement and accuracy. Collaborate asynchronously across time zones with reviewers, project leads, and ML stakeholders.

How To Apply

Apply via Rex.zone and highlight relevant annotation workflows, QA evaluation experience, and examples of guideline-driven labeling that improves training data quality.