Physicist with Python – AI Trainer (Remote, Latin America)

About the role

We are building a talent pool of physics experts with strong Python skills to help evaluate and improve frontier AI models. The role is fully remote and focuses on designing and running physics‑based simulations that test the capabilities of advanced AI systems.

Key responsibilities

• Select and configure a suitable physics simulation package (e.g., PDE solvers, integrators, Monte‑Carlo kernels).
• Create complete Python solutions, including input files, boundary conditions and initial conditions.
• Define numerical outputs and tolerance thresholds that the AI model must achieve.
• Execute the simulations against the AI model in parallel, analyse performance and iteratively adjust difficulty to reach a 10‑30 % pass rate.
• Tune solver parameters and environment settings to probe model behaviour.
• Submit completed tasks to a senior reviewer and incorporate feedback before final delivery.

Required profile

• Academic background in physics (theoretical, experimental or computational) or a related field.
• Minimum two years of hands‑on experience in physics research, applied work or teaching.
• Strong Python programming ability for scientific computing.
• Proven capacity to design problems that require specialized simulation software.
• Excellent written and verbal communication in English.
• Ability to work independently in a fast‑paced remote environment.

Required skills

• Python
• Linux container or sandbox execution environments
• Domain‑specific simulation tools such as FEniCS/DOLFINx, OpenFOAM, Meep, MPB, openEMS, Geant4, PYTHIA8, ROOT/PyROOT, WarpX, REBOUND, MESA, CAMB, CLASS, Bilby

What we offer

• Competitive compensation paid in USD.
• Fully remote work with flexible hours.
• Access to a “Better Me Fund” for personal development.
• Community focused on social good and collaborative learning.