Advanced Math Problems Creation
Budget: $30 – $250 USD
I'm looking for a strong mathematician to author original, university and graduate-level math problems with fully worked solutions. These problems are used to test the reasoning ability of advanced AI models, so they need to be genuinely hard — but still solvable by hand with pen and paper.
Acceptance Criteria:
**I need 1 "very difficult", original tasks. Difficulty can only be gauged by me once I input it in my system where the task will be attempted by multiple Frontier AI models. It will be rated difficult if less than 50% of the models are able to solve the problem. The task has to be delivered in 12 hours of launch.**
What each problem package must include:
1. One original, self-contained problem written in LaTeX (text-only, single question).
2. The correct final answer (uniquely determined).
3. A step-by-step worked solution (minimum 3 substantial steps, fully justified, final answer in \boxed{...}).
4. Domain classification (from a fixed list of 13 math domains + sub-domain) with a short written justification.
5. Problem type, answer type, and a list of 1–5 required concepts.
Hard requirements:
- Must be 100% your own original work — no copying from textbooks, competitions, or existing datasets, and no AI/LLM-generated content of any kind (this is strictly checked and is grounds for rejection).
- Hand-solvable — no calculators, no code, no numbers above 10^15 or below 10^-10, no infeasible computations.
- Not a proof — no "show that / prove / verify" prompts.
- Single unique answer, clearly specified answer format, no multiple-choice or guessable answers, no multi-part questions.
- Clean, correct LaTeX and rigorous, verifiable reasoning in every step.
Domains (pick whatever you're strongest in): Algebra/Functions/Trigonometry, Geometry, Calculus, Linear Algebra, Abstract Algebra, Number Theory, Discrete Math & Combinatorics, Probability & Statistics, Differential Equations & Dynamical Systems, Analysis, Topology & Geometry, Optimization & Numerical Math, Logic/Set Theory/Foundations.
Ideal background: Graduate degree (MSc/PhD or in progress) in mathematics or a closely related field, competition-math experience, comfort with LaTeX, and the ability to write clear, rigorous solutions.
To apply, please include:
1. Your math background / areas of specialization.
2. One short sample problem you authored (with answer + brief worked solution, in LaTeX) at roughly graduate difficulty — this is the main thing I'll evaluate.
3. Your availability and how many problems per week you could produce.
I'll share a detailed brief (formatting rules, domain list, full quality checklist, and example problem profiles) with shortlisted candidates. Applications that ignore the sample-problem request or that look AI-generated will be skipped.
Acceptance Criteria:
**I need 1 "very difficult", original tasks. Difficulty can only be gauged by me once I input it in my system where the task will be attempted by multiple Frontier AI models. It will be rated difficult if less than 50% of the models are able to solve the problem. The task has to be delivered in 12 hours of launch.**
What each problem package must include:
1. One original, self-contained problem written in LaTeX (text-only, single question).
2. The correct final answer (uniquely determined).
3. A step-by-step worked solution (minimum 3 substantial steps, fully justified, final answer in \boxed{...}).
4. Domain classification (from a fixed list of 13 math domains + sub-domain) with a short written justification.
5. Problem type, answer type, and a list of 1–5 required concepts.
Hard requirements:
- Must be 100% your own original work — no copying from textbooks, competitions, or existing datasets, and no AI/LLM-generated content of any kind (this is strictly checked and is grounds for rejection).
- Hand-solvable — no calculators, no code, no numbers above 10^15 or below 10^-10, no infeasible computations.
- Not a proof — no "show that / prove / verify" prompts.
- Single unique answer, clearly specified answer format, no multiple-choice or guessable answers, no multi-part questions.
- Clean, correct LaTeX and rigorous, verifiable reasoning in every step.
Domains (pick whatever you're strongest in): Algebra/Functions/Trigonometry, Geometry, Calculus, Linear Algebra, Abstract Algebra, Number Theory, Discrete Math & Combinatorics, Probability & Statistics, Differential Equations & Dynamical Systems, Analysis, Topology & Geometry, Optimization & Numerical Math, Logic/Set Theory/Foundations.
Ideal background: Graduate degree (MSc/PhD or in progress) in mathematics or a closely related field, competition-math experience, comfort with LaTeX, and the ability to write clear, rigorous solutions.
To apply, please include:
1. Your math background / areas of specialization.
2. One short sample problem you authored (with answer + brief worked solution, in LaTeX) at roughly graduate difficulty — this is the main thing I'll evaluate.
3. Your availability and how many problems per week you could produce.
I'll share a detailed brief (formatting rules, domain list, full quality checklist, and example problem profiles) with shortlisted candidates. Applications that ignore the sample-problem request or that look AI-generated will be skipped.