Physics Annotator (AI)

TL;DR

Physics Annotator (AI): Evaluating AI-generated physics derivations from first principles, checking dimensional analysis, conservation laws, and conceptual integrity with an accent on classical mechanics, quantum field theory, and electrodynamics. Focus on identifying model errors in boundary conditions, approximation regimes, and operator algebra, then authoring rigorous correct solutions as training data for language models.

Location: Fully remote, work on your own schedule from anywhere

Salary: Starting at $40/hour, scales with complexity and performance

Company

AI training data platform connecting physics experts with tasks to improve model reasoning on advanced physics problems.

What you will do

• Craft precise physics prompts to probe AI model reasoning boundaries and expose conceptual flaws.
• Review and evaluate AI-generated solutions line-by-line for dimensional consistency, conservation laws, and physical accuracy.
• Author complete, rigorous correct solutions demonstrating proper reasoning paths for model training.
• Work on projects in classical mechanics, quantum mechanics, electrodynamics, thermodynamics, or general relativity based on your expertise.

Requirements

• Strong physics background: PhD/Master's students, postdocs, research physicists, or applied physicists with graduate-level coursework in QM, E&M, GR, or similar.
• Ability to reason from first principles, identify flawed derivations, and apply physical intuition comparable to peer review.
• Comfortable with analytical mechanics, perturbation theory, Maxwell's equations, partition functions, tensor calculus, etc.
• Pass a short physics assessment (~30 min) focused on reasoning and error identification.
• Commitment to 10+ hours/week around your schedule for consistent high-quality output.

Nice to have

• Published research or peer review experience.
• Industry R&D or simulation fluency.
• Experience in lab/theory groups.

Culture & Benefits

• Flexible scheduling: no fixed hours, standups, or managers; tasks available 24/7.
• Per-task payment processed weekly, competitive with adjunct teaching.
• Direct impact on training next-gen language models; your corrections become training signal.
• High standards enforced, focused on quality over quantity.

Hiring process

• Complete short physics assessment to qualify based on reasoning skills.
• Get matched to projects in your strength areas.
• Start working on tasks and receive weekly payouts.