Subject Matter Expert (Physics) - Prompt Creator (US)

DataForce by TransPerfect

39d • Remote

About The Position

DataForce by TransPerfect is seeking doctoral-level experts in STEM disciplines (Physics, Chemistry, and Biology) to contribute to a highly specialized, high-impact AI training initiative. In this role, you will leverage your own doctoral research to craft "stumper" prompts—complex, PhD-level scientific problems specifically designed to be unsolvable by current Large Language Models (LLMs).
You will be responsible for designing the problem, testing it against various LLMs to verify they fail, documenting their flawed outputs, and finally providing the authoritative, step-by-step expert solution. This is an analytically demanding role that requires deep subject-matter expertise, rigorous scientific reasoning, and proficiency with digital documentation tools.
If you can commit 15 hours per week, the project is expected to take about 8 weeks to complete. The average handling times will be shared at a later stage.

Requirements

• A Ph.D. (completed or in the final stages of defense) in one of the following subdomains: Domain: Physics Polymer Physics Basic Physics Condensed Matter Physics Quantum Physics Computational Physics Applied Physics High Energy & Particle Physics Astrophysics
• Prior experience with peer review, academic grading, or scientific publishing.
• Deep expertise in a specific doctoral research niche.
• Familiarity with interacting with GenAI/LLMs and a critical eye for identifying AI hallucinations or logical leaps in scientific reasoning.
• Exceptional analytical and critical thinking skills.
• Proficiency in standardized scientific documentation (e.g., SMILES, InChIKey for Chemistry; standard notation and equations for Physics).
• Comfort working in an agile data collection/annotation platform environment.
• Meticulous attention to detail and a commitment to strictly original work (AI assistance in generating your own solutions is strictly prohibited).

Responsibilities

• Prompt Creation: Design original, highly complex, self-contained STEM problems based on your doctoral research or directly related subfields.
• LLM Testing & Verification: Input your prompts into multiple designated LLMs to verify that they consistently fail to produce the correct answer ("stumping" the models).
• Authoritative Solution Generation: Write a flawless derivation or calculation leading to the correct answer, including all relevant steps, explanations, equations, and visual diagrams.
• Platform Navigation: Utilize our proprietary DataForce Platform (DFP) to draft, test, and format your submissions (e.g., drawing chemical structures, creating physics diagrams, utilizing equation editors).
• Peer Review (For Senior/QA roles): Top-performing experts will be selected to review peer submissions, ensuring scientific accuracy, clarity, and adherence to project guidelines.

What This Job Offers

• Education Level: Ph.D. or professional degree