Remote | Statistical Computing & Applied Mathematics Expert — $60–$80/hour

24-Mag•New York, NY

3d•Remote

About The Position

We are sharing a specialised part-time consulting opportunity for experienced statisticians and applied mathematicians with graduate-level expertise in computational methods, numerical modelling, and specialised scientific software. This role supports research into advanced computational problem solving. Selected experts will design original graduate-level problems based on realistic statistical and mathematical workflows, develop reproducible reference solutions, test problem difficulty, and create tasks requiring sophisticated use of R, Python, MATLAB, Scilab, and other specialised computational tools.

Requirements

Master's degree or PhD in Statistics, Applied Mathematics , or a closely related quantitative field
PhD preferred, or a Master's degree combined with substantial relevant professional or research experience
Deep hands-on expertise with at least one specialised statistical, mathematical, or scientific software package
Demonstrated computational work through research publications, professional projects, or open-source contributions
Strong Python programming skills
Strong experience with R , MATLAB, Scilab, or comparable numerical computing environments
Practical understanding of numerical methods, modelling assumptions, convergence, diagnostics, and computational limitations
Ability to design rigorous quantitative problems and independently verify solutions
Comfortable working in Linux and terminal-based environments
Strong written communication and ability to explain complex quantitative reasoning clearly

Nice To Haves

Experience with scientific teaching, problem-set design, computational reproducibility, or structured evaluation is advantageous

Responsibilities

Create original graduate-level problems in statistics, applied mathematics, and related quantitative disciplines
Develop tasks requiring multi-step computational reasoning rather than straightforward formula application
Design reproducible problems with clearly defined inputs, outputs, and validation criteria
Create scenarios requiring strategic experimentation, numerical investigation, or inference from partial results
Refine problem designs based on testing and feedback
Develop problems involving Bayesian statistics, advanced regression, latent-variable models, and statistical learning
Design workflows involving parameter estimation, model comparison, diagnostics, and uncertainty quantification
Evaluate whether computational methods are appropriate for the statistical problem being solved
Incorporate realistic edge cases and numerical limitations
Create computational tasks involving time-series modelling, state-space methods, and stochastic processes
Work with survival and event-history analysis using packages such as survival, flexsurv, timereg , and mets
Design differential-equation and dynamical-system workflows using tools such as deSolve, pomp , and FME
Evaluate numerical stability, modelling assumptions, and interpretation of results
Develop problems in spatial statistics, geostatistics, and geographic data analysis
Create tasks involving computational geometry, topology, or specialised quantitative analysis
Apply tools such as TDAstats, geometry, deldir , and polyclip where relevant
Design problems that require careful interpretation of multidimensional or spatial results
Create tasks involving optimisation, mathematical programming, and numerical computation
Apply tools such as nloptr, lpSolve, DEoptimR, SQUAREM , or comparable packages
Develop problems involving numerical linear algebra and high-precision computation
Work with libraries such as RSpectra, Rmpfr, gmp , and pracma
Evaluate convergence behaviour, numerical precision, solver selection, and computational efficiency
Write problem setups, oracle functions, and solution validators
Develop reproducible computational workflows using Python and R
Apply MATLAB or Scilab for numerical modelling and scientific computation where relevant
Verify that expected outputs are mathematically and computationally correct
Document assumptions, parameters, dependencies, and validation logic clearly
Test tasks against advanced computational systems
Identify whether problems are too easy, overly ambiguous, or computationally impractical
Refine tasks until they achieve the intended difficulty level
Design challenges where strong reasoning is required to distinguish between multiple plausible approaches
Ensure problems reward genuine quantitative understanding rather than surface-level pattern matching