Remote (Phoenix, AZ)
This role is with Mercor. Mercor uses RippleMatch to find top talent.
Mercor is seeking a highly skilledResearch and STEM Expertto join our AI evaluation and technical quality assurance team. In this role, you will analyze, evaluate, and fact-check AI-generated outputs across scientific, mathematical, and technical domains — ensuring the highest standards of factual accuracy, logical reasoning, and clarity.
You will help improve the reasoning and reliability of cutting-edge Large Language Models (LLMs) by providing structured feedback and expert judgment across diverse STEM fields. This position is ideal for individuals with strong academic training, analytical precision, and a passion for advancing AI alignment in research and science.
Evaluate and critique AI-generated responses in STEM-related subjects (e.g., computer science, mathematics, physics, biology, and engineering).
Conductfact-checkingandresearch validationusing reputable public and academic sources.
Assess scientific explanations, calculations, and reasoning for correctness and clarity.
Provide structured written feedback to improve the model's understanding and communication of technical topics.
Collaborate with the AI quality team to improve annotation guidelines and maintain consistency across evaluations.
BS, MS, or PhDin aSTEM domain(e.g., Computer Science, Mathematics, Biology, Physics, Engineering, etc.)
English expertwith excellent comprehension and communication skills
Excellent at high school–level math
Experts at fact-checkinginformation across multiple domains (medical, legal, financial, technical, etc.) using trusted public sources
Excellent writing skillsand attention to detail
Prior experience withRLHF annotationorAI model evaluation
Research or professional experience involvingdata analysis,technical writing, oranalytical reasoning
Familiarity with academic research standards and citation practices
Type:Part-time (approximately 20 hours/week)
Location:Remote and asynchronous
Position:Contractor role viaMercor
Rate:$90/hour, based on expertise and domain experience
Payments:Weekly viaStripe Connect