AI Math Task Designer & Evaluator

Crossing Hurdles; United States; Remote; CAD 90,000 - 120,000; Full time

Job summary

Crossing Hurdles seeks a researcher to craft realistic computational mathematics and statistics tasks that push AI models. You will design instructions, reproducible environments, reference solutions, and evaluation criteria, then assess model proofs, calculations, code, and optimizations.

The role requires strong Python skills, deep mathematical expertise, and hands-on experience with numerical methods, optimization, and statistical modeling to ensure robust, reproducible benchmarks.

Qualifications

• Prior paid experience training, evaluating, benchmarking, or improving AI systems on mathematical or technical tasks.
• Strong practical proficiency with Python.
• Deep expertise in at least one relevant mathematical specialization.
• Hands‑on experience using computational methods for numerical analysis, optimization, statistics, theorem proving, modeling, or scientific computing.

Responsibilities

• Create realistic computational mathematics and statistics tasks that challenge advanced AI models.
• Develop task instructions, reproducible environments, reference solutions, and evaluation criteria.
• Evaluate model-generated proofs, calculations, code, models, optimizations, and technical deliverables.
• Identify logical errors, invalid assumptions, numerical instability, incorrect statistical methods, or reproducibility failures.

Skills

Python proficiency

Advanced mathematical specialization

Numerical analysis

Modeling

Theorem proving

Education

PhD in Mathematics or Computer Science

Tools

NumPy

SciPy

Optimization libraries

Job description

Crossing Hurdles seeks a researcher to craft realistic computational mathematics and statistics tasks that push AI models. You will design instructions, reproducible environments, reference solutions, and evaluation criteria, then assess model proofs, calculations, code, and optimizations.

The role requires strong Python skills, deep mathematical expertise, and hands-on experience with numerical methods, optimization, and statistical modeling to ensure robust, reproducible benchmarks.