Entry Level Data Labeling Specialist (United States, Remote)

About the role

You will label and evaluate training data that improves AI/ML systems, focusing on consistent guideline adherence and measurable training data quality. Work includes text, images, and multimodal evaluation tasks, plus structured QA workflows.

Responsibilities

• Perform data labeling and data annotation across NLP, computer vision, and multimodal datasets
• Apply annotation guidelines and document edge cases for adjudication
• Execute QA evaluation workflows (spot checks, consensus review, error categorization)
• Support RLHF by rating responses, ranking candidates, and flagging hallucinations or policy risks
• Conduct prompt evaluation and evaluation set curation to improve model performance
• Maintain throughput and accuracy targets while protecting sensitive data

What You Will Work On

• LLM evaluation: helpfulness, harmlessness, honesty, instruction following
• Content safety labeling: policy-based classification, severity rating, escalation signals
• NLP labeling: named entity recognition (NER), topic tagging, intent detection, factuality checks
• Computer vision annotation: bounding boxes, polygons, segmentation masks, image QA

Requirements

• Comfort with detailed rules, annotation rubrics, and high-precision work
• Strong written communication for documenting rationales and disagreements
• Ability to learn new labeling tools and taxonomy updates quickly
• Familiarity with annotation concepts (NER, CV annotation, RLHF) is helpful but not required