RLHF Specialist

We are seeking RLHF Specialists to compile preference data that aligns models with human intent. You will review multiple model completions and rank them based on truthfulness, helpfulness, and safety.

What you'll do

• Compare model responses and construct preference pairings (preferred vs. rejected)
• Provide written justifications explaining ranking choices using detailed guides
• Conduct multi-turn dialogue evaluations to judge context retention and coherence
• Assist in calibrating training datasets to minimize model sycophancy and bias

What we're looking for

• Exceptional critical thinking skills and cognitive empathy
• Background in philosophy, ethics, cognitive science, or communication
• Proven ability to adhere to highly specific and abstract guidelines
• Prior experience in AI safety or alignment work is highly valued

Benefits & Perks

• Flexible schedule with contract autonomy
• Fully remote collaboration model
• Be at the forefront of safe AI alignment methodologies
• Opportunity for continuous learning and professional development