AI Benchmark Designer: Computational Statistics

Obsidian; Toronto; On-site; CAD 90,000 - 120,000; Full time

Job summary

Obsidian is seeking a Computational Statistics and Applied Mathematics Expert to design challenging computational problems for AI benchmarking. You will craft setups requiring advanced statistical software, run simulations, and interpret results within real scientific workflows.

Ideal candidates hold a graduate degree with hands-on experience in specialized packages and Python; remote-friendly collaboration is encouraged for a time- and task-based engagement.

Qualifications

• Graduate-level training in statistics, applied mathematics, or related STEM field.
• Proven proficiency with at least one specialized software package evidenced by publications or work.
• Strong Python skills for problem setups, validators, and experimentation.

Responsibilities

• Design graduate-level computational problems using specialized software.
• Develop reproducible numerical workflows and test against AI models.
• Refine problems based on feedback and model performance.
• Collaborate with domain experts to ensure realistic scientific workflows.

Skills

Bayesian stats

IRT/psychometrics

SEM/latent models

Topological data analysis

Differential equations

State-space/time-series

Survival analysis

Mixed models

Spatial statistics

Statistical learning

Optimization

Numerical linear algebra

Computational geometry

Education

MS or PhD in statistics or applied mathematics

Tools

rstan

cmdstanr

rjags

brms

TAM

sirt

mirt

lavaan

OpenMx

TDAstats

pomp

forecast

vars

rugarch

statsmodels

PyMC

Job description

Obsidian is seeking a Computational Statistics and Applied Mathematics Expert to design challenging computational problems for AI benchmarking. You will craft setups requiring advanced statistical software, run simulations, and interpret results within real scientific workflows.

Ideal candidates hold a graduate degree with hands-on experience in specialized packages and Python; remote-friendly collaboration is encouraged for a time- and task-based engagement.