Senior Physics AI Researcher (Remote)

Synthires; United States; Remote; USD 124,000 - 193,000; Part time

Job summary

Synthires is seeking a Physics Domain Expert to work remotely from the United States on an hourly contract. You will review AI-produced physics work, craft rigorous reference solutions, and design benchmarks to advance scientific reasoning in frontier AI models.

The role requires a PhD in physics or related field, senior research experience, and deep specialization in a physics domain, with hands-on evaluation of language models.

Qualifications

• PhD in Physics, Astrophysics, Applied Physics, or a closely related field.
• Senior research experience as a faculty member, PI, staff or senior scientist, or equivalent.
• Deep specialization in at least one physics domain such as condensed matter, particle physics, astrophysics, cosmology, quantum physics, or computational physics.
• Hands-on experience using large language models and evaluating scientific reasoning.
• Excellent written communication and analytical skills.
• Must be based in the United States.

Responsibilities

• Review physics tasks and AI-generated outputs for scientific accuracy, flawed derivations, and unsupported assumptions.
• Write precise problem specifications and rigorous reference solutions.
• Develop research-level problems and benchmarks in your area of specialization.
• Collaborate with researchers to establish clear evaluation criteria.
• Provide structured feedback on the quality and reliability of physics reasoning.

Skills

LLM experience

Written communication

Analytical thinking

Education

PhD in Physics or related field

Job description

Synthires is seeking a Physics Domain Expert to work remotely from the United States on an hourly contract. You will review AI-produced physics work, craft rigorous reference solutions, and design benchmarks to advance scientific reasoning in frontier AI models.

The role requires a PhD in physics or related field, senior research experience, and deep specialization in a physics domain, with hands-on evaluation of language models.