Overview

In this hourly-paid, fully remote contractor role, you will support the development of advanced AI safety systems by reviewing AI-generated responses and/or creating safety-focused evaluation content. You will assess reasoning quality, step-by-step problem-solving, and overall response safety, providing structured, high-quality feedback to improve model accuracy, reliability, and alignment. You will annotate and evaluate content in both English and Arabic, ensuring linguistic precision and cultural nuance in safety-critical contexts. Your evaluations will directly contribute to preventing large language models from generating unsafe, biased, or adversarial outputs, including content related to hate, violence, sexual material, self-harm, misinformation, and other sensitive domains. This role is part of a fast-growing AI data services organization that provides high-quality training data to leading AI companies and foundation model labs, helping shape the safety a...