Physics Expert for AI Model Evaluation

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We connect skilled contributors with projects that shape how AI systems learn by providing clear project details, transparent pay, and tools to grow a durable freelance career in AI training.

About AI training work

AI training (data labeling and human feedback) is the human foundation of modern AI: people create, review, and judge examples that teach models to reason, write, and solve technical problems. This work is remote, often flexible and part-time, and lets domain experts directly influence model behavior and reliability.

• Work is typically remote and flexible — fit it around academic or research commitments.
• Projects range from annotation to high-level model evaluation and RLHF-style feedback.
• Specialist expertise (like advanced physics) is highly valued and commands higher pay.

The role

We are recruiting senior physics experts — established academics and research leaders — to provide high-level domain guidance for training advanced AI systems. You will deliver rigorous, defensible judgments about physics reasoning, compare approaches, and help define evaluation criteria that improve model behavior and scientific reliability.

Labeling work for this project includes evaluation ratings and writing prompt/response examples (SFT-style) to teach models how to reason about physics problems.

• Project type: remote, contract, part-time work focused on expert evaluation and response writing.
• Labeling tasks include: EVALUATION_RATING and PROMPT_RESPONSE_WRITING_SFT.

What you'll do

• Adjudicate contested physics arguments, solutions, and interpretations within your subfield.
• Compare alternative problem-solving approaches and explain when one is superior under stated assumptions.
• Identify and define evaluation criteria for robustness and validity of physics reasoning.
• Provide authoritative written assessments while clearly noting genuine uncertainties or open questions.
• Draft rigorous, defensible written evaluations for review by senior physicists.
• Use technical tools (LaTeX, SymPy, Python, Jupyter) to verify or contrast technical claims.
• Clearly communicate when a question is unresolved and outline pertinent considerations.

Requirements

This role requires a strong, verifiable academic and leadership record in physics and hands-on technical familiarity with common research tools.

• PhD in physics with demonstrated scholarly impact in your specified subfield.
• Current or former Associate Professor, Full Professor, Chair Professor, or Principal Investigator/Group Leader with independent research leadership.
• Active research in one or more areas listed: High Energy Physics, Mathematical Physics, Biophysics, Statistical Physics, Condensed Matter, AMO/Quantum Optics, Gravitation, Cosmology, Astrophysics, Quantum Information, or Optical Properties of Materials.
• 3–5 recent representative publications in your target subfield with arXiv links or DOIs.
• Prior experience supervising PhD students or postdocs, or equivalent industry research leadership.
• Proficiency with LaTeX, SymPy, Python, and Jupyter.
• Exceptional written communication skills, able to articulate nuanced judgments clearly in English.

Workload, schedule, and pay

This project requires approximately 10 hours per week over an 8–10 week window. The schedule is flexible and fully remote; contributors may be located anywhere worldwide and work asynchronously.

Employment type: contract, part-time. Payment is hourly in USD.

• Time: ~10 hrs/week; total duration 8–10 weeks (fits under Less than 20 hours/week).
• Pay: hourly, USD — range $80–$160/hr (listed rate $160/hr).
• Location: remote, worldwide. Language: English.

How it works and how to apply

Apply through OpenTrain with your CV and evidence of the qualifications listed below. We will review submissions for research leadership, publications, and tool proficiency before inviting candidates to complete a short sample evaluation to confirm fit.

Successful applicants will receive clear instructions, task guidelines, and project milestones. Work is reviewed by senior subject-matter raters to ensure quality and consistency.

• Prepare: CV, 3–5 representative publications (arXiv/DOI), brief notes on supervisory experience and current research activity.
• Expect to demonstrate proficiency with LaTeX, Python/Jupyter, and symbolic tools in a short sample task.
• OpenTrain provides transparent project details and support throughout the engagement.