Job description### Computational Statistics and Applied Mathematics Expert ### About the Project We're building a large-scale benchmark to test how well advanced AI systems can solve hard scientific and engineering problems. As a task designer, you'll create challenging computational problems that check whether AI can use real scientific software to do research-level work — running simulations, interpreting results, designing experiments, and uncovering hidden information from data. This isn't a typical data-labeling job. You'll design original, graduate-level problems based on real scientific workflows, test them against cutting-edge AI models, and fine-tune them until the difficulty is just right. ### What You'll Do You'll create problems that require skilled use of specialized statistical, mathematical, or scientific software packages. Some will ask the AI to compute reproducible numerical answers from a fully defined setup — testing whether it can correctly carry out complex, multi-step workflows. Others will be harder: the AI must plan a series of queries or experiments to uncover information that isn't directly visible, which means thinking strategically about what to measure, how to read partial results, and how to narrow down the possibilities efficiently. Each problem goes through a testing loop against state-of-the-art AI models, and you'll refine it until it hits the target difficulty. ### Domains & Tools We're Hiring For We welcome statisticians and applied mathematicians working across a wide range of specializations. You do not need experience with every package listed below; strong expertise with one or more specialized computational packages is sufficient. We're especially interested in experts with deep, hands-on experience using one or more specialized R or Python packages, including examples such as: - **Bayesian statistics:** rstan, cmdstanr, rjags, runjags, brms, rstanarm, nimble, bayesplot, posterior, loo - **Item response theory and psychometrics:** TAM, sirt, mirt, mirtCAT, eRm, ltm, lordif, psych - **Structural equation and latent variable modelling:** lavaan, semTools, OpenMx - **Topological data analysis:** TDAstats, TDApplied - **Differential equations and dynamical systems:** deSolve, pomp, FME - **State-space and time-series modelling:** KFAS, MARSS, forecast, vars, urca, rugarch, rmgarch, tseries, timeSeries - **Survival and event-history analysis:** survival, flexsurv, timereg, mets - **Mixed, additive, and advanced regression models:** lme4, nlme, mgcv, glmmTMB, TMB, quantreg, scam - **Spatial statistics and geostatistics:** spatstat, spatstat.geom, spatstat.linnet, spdep, gstat, geoR, spBayes, sf, stars, terra, lwgeom - **Statistical learning and specialized modelling:** mclust, kernlab, earth, pROC, multcomp, sandwich, effectsize, irr - **Optimization and mathematical programming:** lpSolve, linprog, nloptr, DEoptimR, SQUAREM - **Numerical linear algebra and high-precision computation:** RSpectra, Rmpfr, gmp, pracma - **Computational geometry:** geometry, deldir, polyclip Other similar specialized statistical, mathematical, scientific, or domain-specific R packages will also be considered. Other similar specialized statistical or mathematical Python/Scilab packages are also welcome, such as statsmodels and PyMC. Numerical computing and scientific modelling in Matlab/Scilab are also wanted. ### What Makes a Strong Candidate You have graduate-level expertise (MS or PhD required; PhD preferred, or MS with 10+ years of relevant experience) in statistics, applied mathematics, or a closely related quantitative field, with real hands-on experience using specialized computational packages — not just theoretical knowledge. You have written code using one or more specialized statistical, mathematical, or scientific packages to solve actual research or professional problems, and you understand where these tools break, what their edge cases are, and what makes a problem genuinely hard rather than just complicated. Deep expertise with one or more specialized computational packages is more important than familiarity with the entire package list above. Beyond domain expertise, the best candidates think like puzzle designers: building problems where the challenge comes from smart reasoning rather than raw computation, where several approaches seem plausible but only careful analysis reveals the right one, and where surface-level pattern matching won't get you to the answer. ### Requirements - Graduate-level training in statistics, applied mathematics, a relevant STEM field, or equivalent research experience - Proven proficiency with at least one specialized statistical, mathematical, or scientific software package, demonstrated through research publications, open-source contributions, or professional work - Strong Python skills — you'll be writing problem setups, oracle functions, and solution validators - Ability to work independently and refine problem designs based on feedback - Comfortable working in a Linux/terminal environment with remote compute sandboxes - Available for at least 15–20 hours per week ### **Nice to Have** - Experience across multiple computational domains or specialized software packages - Familiarity with benchmark or evaluation design - Background in scientific teaching or exam/problem-set design - Experience with computational reproducibility and containerized environments
