AI Safety Practitioner

About the Role

As an AI Safety Practitioner, you will be responsible for assessing the safety, quality, and alignment of advanced AI models, specifically focusing on complex and sensitive topics. Your evaluations will involve AI-generated content, ensuring it meets stringent safety and policy standards, and contributing to the enhancement of model behaviors through structured feedback.

Key Responsibilities

• Assess AI outputs for safety, factual correctness, adherence to policies, and overall quality standards.
• Examine content related to misinformation, political bias, self-harm, violence, cybersecurity, biosecurity, and other critical sensitive areas.
• Utilize and improve evaluation frameworks for reinforcement learning with human feedback (RLHF), supervised fine-tuning (SFT), and AI safety benchmarking.
• Detect and report unsafe content, hallucinations, errors in reasoning, and policy breaches.
• Deliver detailed feedback aimed at enhancing model alignment and safety.
• Work alongside AI researchers and safety teams to support ongoing evaluation projects.

Candidate Profile

• Holds a bachelor's degree or higher in disciplines such as Journalism, Communications, Psychology, Sociology, Public Policy, Law, Biology, Chemistry, Computer Science, or related fields.
• Possesses over five years of professional experience in AI Safety, Trust & Safety, journalism, government policy, scientific research, security, or comparable industries.
• Exhibits strong written English communication, critical analysis, and logical reasoning capabilities.
• Demonstrates consistent proficiency in assessing nuanced and policy-sensitive situations.

Preferred Qualifications

• Prior involvement in AI safety, reinforcement learning with human feedback, supervised fine-tuning, Trust & Safety, or AI evaluation roles.
• Knowledge of safety policies, content moderation, and developing evaluation rubrics.
• Experience handling and reviewing complex, ambiguous, or high-risk content.

Additional Information

This role is a contract position based onsite in the United States. The hourly compensation ranges from $60 to $70. Candidates are expected to have a strong background in AI ethics, safety evaluations, and policy interpretation related to advanced AI systems.