Physicist, PhD

Physics training data

Develop and review datasets that represent physics across topics and levels of difficulty. Work includes problems, concepts, diagrams, experiments, and other scientific material.

• Classify and label physics content accurately.
• Connect principles across mechanics, electromagnetism, thermodynamics, and related subjects.
• Verify generated solutions, identify scientific mistakes, and explain corrections through clear feedback.
• Help create balanced datasets covering varied physics questions and problem complexity.

Required expertise

A doctorate in Physics or a closely related field is required, with demonstrated depth in research, teaching, or advanced problem solving. You should have developed or critically reviewed sophisticated scientific, technical, or educational content.

Relevant experience may include research papers, curriculum design, or complex simulations. Clear communication of advanced physical concepts and the ability to connect abstract theory to practical applications are important.

Data labeling, RLHF, and other AI training or model evaluation experience are advantages.

Contract arrangement

The hourly, at-will engagement is fully remote and asynchronous for at least 15 hours weekly. US-based contributors are preferred, with a US hourly range of $25 to $100.

Starts are arranged as projects open. Contributors have clear responsibility for their work and autonomy in delivery, while strong performers may lead programs or mentor incoming specialists.

Selection

An English-language AI screening covers your background and experience. A subject assessment then explores advanced physics principles, the analysis of complex processes, and applying theory to practical scenarios.

Candidates are reviewed on a rolling basis and contacted when their skills align with an active project.