[Remote] Bayesian Data Scientist
Note: The job is a remote job and is open to candidates in USA. Metasys Technologies is hiring a Research Consultant - Modeling and Quantitative Analysis to support Digital Planet’s computational model and open-access platform on AI readiness in low- and middle-income countries. The role focuses on Bayesian probabilistic modeling, indicator harmonization, diagnostic analysis, reproducible data pipelines, and clear documentation for public release.
Responsibilities
- Work on specification, estimation, and testing of the probabilistic segmentation model, including diagnostics on calibration and cross-country consistency
- Investigate anomalies in model output and propose specification fixes, with the analysis to justify them
- Derive and test the weighting scheme that combines indicators into segment assignments, with weights estimated from data rather than assigned by judgment wherever possible
- Build and maintain reproducible pipelines to clean and harmonize indicator data from a range of international and national sources
- Reconcile indicators across differing vintages, definitions, and geographic levels; document assumptions and source hierarchies
- Produce diagnostic tables, charts, and sensitivity analyses for internal and external review
- Write clear code documentation and methodology notes suitable for public release
Skills
- • Able to interrogate a modeling choice or a data source rather than accept it
- • Able to say plainly when something looks wrong
- • Can explain a method, a diagnostic result, or a caveat clearly, in writing and in conversation
- • Comfortable communicating with both technical and non-technical audiences
- • Comfortable reasoning about probability, uncertainty, and distributions
- • Not limited to running estimation code
- • Prior work specifying and estimating a model of similar structure
- • Experience deriving or validating a weighting scheme
- • Measurement or latent-trait index models, composite indicators, and population segmentation or classification models are considered close analogues
- • Experience with hierarchical and latent-variable specifications
- • Knowledge of prior specification, partial pooling, posterior diagnostics, and convergence diagnostics
- • Ability to judge how much structure a small number of units can support
- • Understanding that with 27 countries, priors and shrinkage carry substantial weight
- • Experience with panel methods built for large-N asymptotics transfers only partly
- • Experience working with indicator coverage that differs by country and year
- • Understanding that data gaps are not random
- • Ability to determine when to: • Impute, • Marginalize, • Let the measurement model absorb missingness
- • Rigorous about units, vintages, provenance, and documentation
- • Willing to learn the project's framework, indicators, and country context
- • Comfortable with a ramp-up period
- • Fluent in Python for data and modeling work
- • Experience with: • pandas • NumPy • SciPy
- • Comfortable using a version-controlled repository
- • Experience with probabilistic programming frameworks such as: • PyMC • NumPyro • Stan
- • Able to accommodate new series and revised definitions without rebuilding specifications and code each time
- • Able to pick up an existing codebase
- • Can make progress with limited direction
- • Raises blockers early
- • Demonstrated experience specifying, estimating, and diagnosing Bayesian hierarchical or latent-variable models, evidenced by prior projects, publications, or code
- • Prior work building at least one comparable model end to end: • Measurement or latent-trait index, • Composite indicator, • Population segmentation or classification model
- • Experience deriving or validating a weighting scheme
- • Experience working with multi-country or multi-year panel data with incomplete or uneven coverage
- • Python for statistical and data work, including: • pandas • NumPy • SciPy
- • Experience with a probabilistic programming framework such as: • PyMC • NumPyro • Stan
- • Comfortable working in a version-controlled repository (Git)
- • Able to pick up an existing codebase and work productively with limited supervision
- • Strong written communication skills
- • Able to clearly document methods, diagnostics, and caveats for both technical and non-technical audiences
- • Able to find and synthesize relevant literature to inform methodological or measurement decisions
- • Brings a fresh perspective to problem-solving within an established framework
- • Experience with dynamic latent trait or cross-national measurement models used to build governance, democracy, or human rights indices
- • Familiarity with household survey programs such as: • DHS • LSMS • Global Findex
- • Experience with sub-national administrative data
- • Graduate training or professional background in: • Statistics • Economics • Data Science • Demography • Related quantitative fields
- • Interest in digital access and AI readiness in low- and middle-income country contexts
- • Experience with dynamic latent trait or cross-national measurement models, such as those used to build governance, democracy, or human rights indices
- • Familiarity with household survey programs such as: • DHS • LSMS • Global Findex
- • Experience with sub-national administrative or geospatial data
- • Experience structuring model code so new indicators and revised definitions can be added without a rebuild
- • Ability to find and synthesize methodological literature to inform a measurement decision
- • Graduate training or professional background in: • Statistics • Economics • Data Science • Demography • Related quantitative fields
- • Interest in digital access and AI readiness in low- and middle-income country contexts
Benefits
- Fully remote within the United States.
- Flexible hours week to week.
Company Overview