Remote Data Annotation Jobs in Rome

About The Role

Remote data annotation jobs in Rome at Rex.zone focus on creating and evaluating high-quality training data for AI systems. In this full-time remote role, you will label text, images, audio, and video; apply detailed annotation guidelines; and perform QA evaluation to improve model performance across NLP, computer vision, and LLM training pipelines.

Key Responsibilities

• Label data across NLP and computer vision tasks (e.g., classification, NER, bounding boxes/segmentation, OCR verification where applicable)
• Perform named entity recognition and relation extraction
• Support RLHF workflows via preference labeling, prompt evaluation, and response ranking for LLM evaluation
• Execute QA evaluation using rubrics, audit sampling, calibration tasks, and inter-annotator agreement checks
• Complete content safety labeling (e.g., harassment, self-harm, adult content, misinformation) as required
• Maintain strict annotation guidelines compliance, document edge cases, and escalate ambiguities
  
  

Required Qualifications

• Mid-senior experience in data annotation, data labeling, and/or QA evaluation for AI/ML projects
• Ability to interpret complex rubrics and maintain high consistency on nuanced edge cases
• Familiarity with LLM evaluation concepts, RLHF-style preference labeling, and structured labeling for NLP (NER) or CV tasks
• Strong attention to detail, time management, and ability to work independently in a remote environment
  
  

Work Model

This is a full-time Remote role with asynchronous collaboration via Rex.zone workflows, including documented decisions, calibration examples, and QA feedback loops to reduce guideline drift.

How To Apply

Apply via Rex.zone with a brief summary of relevant experience and examples of tasks completed (NER, classification, RLHF preference ranking, computer vision annotation, content safety labeling) and the quality methods used to maintain training data quality.

FAQ: Remote work is fully remote; projects may include RLHF/LLM evaluation; QA includes rubric-based reviews and agreement checks.