Red-Teaming QA Lead - Remote

The position is a contract, hourly role performed remotely, where you will act as the lead for quality assurance in AI red‑team and safety‑evaluation initiatives. The employer is a rapidly expanding AI data‑services firm that supplies training material to many of the world’s leading artificial‑intelligence developers and foundation‑model laboratories, aiming to make their models safer through realistic, policy‑aligned data. Each day you will examine AI‑generated safety assessments, adversarial prompts, risk reports and the work of trainers and reviewers, checking them against detailed project rubrics. You will deliver precise written comments, flag recurring problems, and raise critical issues when needed. Communication with the broader team will happen mainly through Discord, where you will post guideline updates, answer questions, and encourage inactive contributors to re‑engage. You will also create and upkeep documentation such as style guides, FAQs, trackers and onboarding resources, and you will lead training sessions that explain expectations and review standards. The role requires a bachelor’s, master’s or comparable professional background in fields such as computer science, cybersecurity, AI safety, public policy, psychology, linguistics, law or related disciplines. You must possess strong command of English for both following instructions and providing clear feedback, and you need at least three years of experience in areas like AI safety, red‑team operations, cybersecurity, trust‑and‑safety, content policy, risk analysis or comparable work. A solid understanding of AI risk types, adversarial prompting, jailbreak tactics, harmful‑content taxonomies and policy interpretation is essential, as is the ability to evaluate work against complex rubrics with meticulous attention to detail. Additional advantages include familiarity with topics such as prompt injection, social engineering, fraud, self‑harm, extremist material, misinformation, privacy concerns, bias and model refusal behavior. Prior experience managing remote groups of red‑teamers, reviewers, policy analysts or quality assessors is highly valued, as is a background in AI training data, large‑language‑model evaluation, safety testing or rubric‑based review. Comfort with fast‑paced remote collaboration tools like Discord, Google Sheets, Docs and project‑tracking dashboards will help you succeed in this dynamic environment.