High-Energy & Mathematical Physics Expert

About OpenTrain

OpenTrain is the #1 platform for building careers in AI training and data labeling. We connect domain experts with projects that shape how AI systems learn and behave, and we help contributors consolidate work, build a portfolio, and grow a durable freelance career in this fast-growing industry.

About AI training work

AI training is the human side of building intelligent systems: people provide the domain knowledge, judgments, and examples that let models learn to reason and respond correctly. This work is largely remote, flexible, and accessible—contributors often work part-time and choose schedules that fit their lives.

For this role, your deep physics expertise will directly influence model behavior by providing high-quality, rigorously argued evaluations and adjudications on technical physics claims.

The role

You will act as a senior subject-matter adjudicator: comparing competing solutions, rendering expert judgments, and drafting rigorous written evaluations that explain when one approach is superior, under what assumptions, and where uncertainties remain.

• Contractor, part-time role: less than 20 hours per week.
• Work is remote and paid hourly at $80–$160/hour (OpenTrain AI hires directly).
• Primary data type: technical text; label type: evaluation/rating.

What you'll do

Daily tasks center on careful, defendable expert judgment and clear technical writing. You will use formal tools as needed to verify or contrast claims, and you will communicate where open questions persist.

• Adjudicate contested or competing physics arguments, solutions, and interpretations in high-energy or mathematical physics.
• Compare alternative approaches and state which is superior, the assumptions required, and the validity regimes.
• Identify meta-level criteria for robustness and validity of competing work.
• Provide calibrated confidence: state authoritative assessments while transparently acknowledging uncertainty or open problems.
• Draft defensible written evaluations suitable for review by senior physicists.
• Use LaTeX, SymPy, Python, and Jupyter to verify or contrast technical claims when needed.
• Clearly flag unresolved questions and enumerate the relevant considerations.

Requirements

We require the specific scholarly and technical qualifications below—please do not apply unless you meet them.

• PhD in physics with demonstrated expertise and scholarly impact in high-energy or mathematical physics.
• Current or former Associate Professor, Full Professor, Chair Professor, or Principal Investigator/Group Leader with independent research leadership.
• Ongoing research activity using one or more of: AdS/CFT, onshell gravitational action, Weyl-Fefferman-Graham gauge, or tensor-algebra of boundary curvature invariants.
• Proficiency with LaTeX, SymPy, Python, and Jupyter.
• Exceptional written communication skills capable of articulating nuanced, well-reasoned technical judgments.

Helpful background and who should apply

Competitive applicants typically can point to recent representative publications and demonstrated leadership in research or editorial roles. Prior AI experience is not required—your domain expertise is what matters.

• 3–5 recent representative publications (arXiv or DOI references) are helpful.
• Experience supervising PhD students or postdocs, or equivalent research leadership, is desirable.
• Service on editorial boards or program committees is a plus.

How the engagement works

OpenTrain AI is the contracting organization. Assignments will consist of short adjudication or evaluation tasks requiring rigorous written answers and occasional technical checks with standard tools. You will be paid hourly within the stated range and can accept work to fit under the stated weekly time limit.

• Apply with CV and (strongly recommended) 3–5 representative publication references (arXiv or DOI).
• Work remotely as a contractor and choose tasks consistent with your available hours (under 20 hrs/week).
• Expect tasks to involve detailed written evaluations and occasional reproducibility checks using Jupyter/Python/SymPy/LaTeX.