AI training jobs for physicists

Write original physics problems with complete worked solutions and evaluate AI-generated derivations and numerical answers for wrong assumptions and errors.

Author physics computing tasks for AI models, each with a main problem, Python reference solution, and tests.

Write physics problems with reference answers and grade model-written derivations and numerical work.

Write benign, dual-use, and adversarial prompts and evaluate model responses against a policy standard, writing reference answers.

Write benign, dual-use, and adversarial prompts and evaluate AI model responses against a defined policy standard.

Analyze statistical physics problems and deliver documented solutions or critiques involving Kramers-Wannier duality and quenched disorder averaging.

Solve advanced physics problems with rigorous derivations written in LaTeX, and verify results using SymPy, Python, and Jupyter.

Review ab initio configuration interaction methods for atomic polarizability and critique atomic-structure calculations for Yb isotopes.

Author original multiple-choice physics questions and review pre-written questions for accuracy, clarity, and rigor.

Review AI-generated physics content and trainer work against project guidelines, and provide written feedback on scientific accuracy.

Build realistic tasks from your own physics work and judge frontier AI model answers.

Classify user requests as benign, dual-use or adversarial, judge three model answers safe or unsafe, and rank them.

Review AI-generated physics responses, evaluate step-by-step reasoning, and identify errors in calculations, methodology, or conceptual understanding.

Train and evaluate AI models in physics, and create tasks and deliverables based on real-world scenarios.

Craft and review physics-related reasoning problems and verify the logical consistency of model responses.