Lead and quality-assure Moonshot's applied AI safety work across harm categories including pathways to violence, extremism, CSEA, abuse and grooming, mental health and crisis, and risks affecting children and teens, using methods such as red teaming and adversarial evaluation of AI systems. This is not an engineering or data-science role, but it is a hands-on position requiring the successful candidate to lead and participate directly in red teaming and adversarial evaluation, working in detail with evaluation methodologies, test scenarios, model responses, safety policies, and intervention frameworks.