AI Trainer - Neuroscience; Fully Remote- St Louis

Role overview

This remote advisory role invites neuroscience specialists to evaluate graphical abstracts produced by large language models against their underlying scientific papers. The work supports an AI verification effort aimed at preserving scientific integrity and factual precision in machine-generated technical summaries. It is a flexible, ad hoc project suited to researchers looking to apply domain expertise on a part-time basis.

• Read peer-reviewed neuroscience papers and compare them side by side with LLM-generated graphical abstracts.
• Identify factual errors, misrepresentations, and structural flaws in the AI-produced visual summaries.
• Verify that technical terminology, methods, and findings are depicted accurately in both imagery and text.
• Flag discrepancies and provide clear written feedback on where abstracts diverge from the source research.
• Pass an initial paid skills assessment using a sample paper and abstract before ongoing assignments.

• Master's or PhD in neuroscience or a closely related discipline.
• Deep familiarity with experimental methods, terminology, and conventions in neuroscience literature.
• Demonstrated ability to read and critique dense scientific publications with care.
• Proven fact-checking experience or a track record of detailed analytical review work.
• Strong attention to detail and comfort producing structured, evidence-based judgments.

• Prior involvement in AI evaluation, training data annotation, or reinforcement learning with human feedback.
• Background in scientific publishing, peer review, or journal editorial work.

• Fully remote with flexible hours.
• Hourly compensation starting at $40+, paid for the qualifying skills assessment as well as project work.