Location: Lemont, IL
Cap-Exempt H-1B Position — No lottery required
Join Argonne National Laboratory’s multidisciplinary biomedical data science team to contribute to cutting-edge research at the intersection of artificial intelligence, predictive health modeling, and translational biomedical analytics. This predoctoral appointment offers an opportunity to work on large-scale validation of advanced predictive models derived from harmonized longitudinal human datasets, with applications in long-term health outcome prediction and proactive healthcare decision support. The selected candidate will help evaluate scientific rigor, reproducibility, robustness, and generalizability of computational models in a collaborative environment that integrates data science, biostatistics, and biomedical research. Core Responsibilities: - Support reproducibility studies of predictive machine learning models by independently executing analytical pipelines on harmonized datasets and verifying reported performance metrics. - Conduct external validation of predictive models using independent datasets to assess model generalizability across populations and settings. - Develop and maintain reproducible computational workflows for secure execution of data pipelines and model benchmarking. - Perform preprocessing, normalization, feature harmonization, and quality control on large longitudinal biomedical datasets. - Conduct sensitivity analyses to evaluate model robustness under input perturbations, parameter variation, and missing-data scenarios. - Assess potential bias introduced by data imputation and harmonization methods in long-horizon predictive modeling. - Generate statistical analyses, benchmarking summaries, visualizations, and technical documentation for internal and external reporting. - Collaborate with interdisciplinary teams including computational scientists, statisticians, software engineers, and domain researchers to improve model validation methodologies. Learning Opportunities: - Gain hands-on experience in independent validation of AI mod