Data Annotation Tech, Outlier/ScaleAI – Remote
• Label, classify, and annotate data (text, images, and structured data formats [JSON]) as required. • Create and script prompts (Adversarial, Domain-Specific & Edge Cases) for various purposes, including testing the model's behavior. • Create red-team prompts specifically designed to test the model for hallucinations (factual inaccuracies) and failures to follow instructions. • Create prompts for data gathering and generation across various domains (e.g., programming, creative writing, conversational artificial intelligence). • Review AI-generated output for accuracy, coherence, relevance, and completeness in order to enhance the quality of training datasets. • Complete “Rate & Review” assignments to evaluate AI-generated output against defined rubrics and assign a level of quality, accuracy, helpfulness, and safety. • Create & evaluate JSON files for AI training datasets, ensuring compliance with the schema, accuracy, and formatting consistency. • Labeled images using annotation tech