Mathematician & Python AI Trainer (Remote, Latin America)

About the role

We are building a talent pool of mathematics professionals with strong Python skills to evaluate and improve frontier AI models. This remote, project‑based position is ideal for mathematicians who enjoy deep technical problem‑solving and want to apply their expertise to cutting‑edge AI research.

Key responsibilities

• Identify the most suitable mathematical software package and design problems that rely on its core capabilities.
• Develop complete Python solutions for each problem, including any necessary input files.
• Define the correct answer and acceptable margin for AI model responses based on the mathematical context.
• Run problems against AI models in parallel, analyze performance, and adjust difficulty to achieve a 10‑30% pass rate.
• Iteratively refine problem parameters and document how the model handles complex challenges.
• Submit completed tasks to a senior reviewer in the sub‑field and incorporate feedback before final submission.

Required profile

• Academic background in pure or applied mathematics (or equivalent).
• Minimum 2 years of experience in mathematics research, applied work, or teaching.
• Strong Python programming skills for computational solutions.
• Ability to create problems that cannot be solved without specialized mathematical software.
• Excellent written and verbal English communication.
• Self‑motivated and able to work independently in a fast‑paced remote environment.

Required skills

• Python
• Z3, cvc5, SageMath, Macaulay2, Singular, CasADi, IPOPT, SDPB, G6K, fpylll (or ability to learn them quickly)
• Experience with containerized or sandboxed Linux execution environments

What we offer

• Competitive pay in USD
• Fully remote work arrangement
• Better Me Fund for personal growth and passions
• Community focused on social good and compassionate collaboration