Research Specialist

University of Chicago•Hyde Park, IL

1d•$50,000 - $53,000•Onsite

About The Position

We are seeking a Researcher to work at the intersection of AI/ML, particle physics, and medical physics. The successful candidate will have expertise applying AI/ML in at least one of these areas — particle physics or medical physics — and a willingness to learn new methods and domains and to transfer approaches across them. The role centers on developing modular, acquisition-aware AI foundation models and applying them to large-scale scientific and biomedical imaging data, and is well suited to someone excited to work across disciplines and to contribute to open, reproducible computational resources. A primary focus of the position is developing a modular, acquisition-aware, multimodal AI foundation model designed to be generic to diseases of the brain, learning shared representations across conditions rather than being tied to a single disease. The Researcher will help build the model, assemble and harmonize multimodal brain-imaging datasets, and run the benchmarking and cross-condition transfer experiments at the heart of the project. The Researcher will also contribute to the group’s broader program of developing AI/ML methods for particle physics and the modeling of physical systems — including deep-learning approaches to reconstruction, classification, and analysis of large-scale detector data — and will help transfer modeling advances between the physics and brain-health domains. This position provides research and technical support activities related to scientific research projects, and ensures compliance of research activities with institutional, state, and federal regulatory policies, procedures, directives, and mandates. This position is expected to last approximately one year, with the possibility of extension based on funding.

Requirements

Minimum requirements include a college or university degree in related field.
Minimum requirements include knowledge and skills developed through < 2 years of work experience in a related job discipline.

Nice To Haves

Master’s degree in machine learning/AI, physics, applied mathematics, or a related quantitative field.
Demonstrated experience developing AI/ML and deep-learning methods, including model design, training, and evaluation.
Background in medical physics research or biomedical imaging (e.g., MRI, PET, or related modalities), and/or experience developing AI/ML methods for particle physics or other large-scale scientific/detector data.
Experience with neuroimaging data and standard preprocessing tools.
Familiarity with foundation models, transfer learning, self-supervised learning, or knowledge distillation.
Research experience.
Strong programming skills and experience with modern deep-learning frameworks (e.g., PyTorch).
Experience with high-performance or GPU computing environments.
Familiarity with version control, containerization, and reproducible-research tooling.
Work independently and collaboratively within a multidisciplinary research environment.
Communicate methods and results clearly, in writing and in presentations.
Manage multiple tasks and priorities and follow projects through to completion.
Contribute to open-source software and reproducible scientific workflows.

Responsibilities

Designs, implements, and trains components of the foundation-model architecture, including modality-specific encoders, cross-modal fusion, acquisition-aware conditioning, and prediction heads.
Develops and applies self-supervised training objectives (e.g., masked signal modeling, cross-modal contrastive learning) and knowledge-distillation strategies from existing foundation models.
Curates, preprocesses, and harmonizes multimodal data (structural and functional MRI, PET, EEG, and clinical/phenotypic measures) from established repositories such as ADNI, OASIS-3, NACC, and BSNIP, including acquisition-metadata handling and cross-site harmonization.
Builds and runs rigorous benchmarking evaluations comparing the model against disease-specific, single-modality, and existing foundation-model baselines, including pre-specified metrics, ablation studies, and cross-cohort generalization tests.
Implements and evaluates zero-shot and few-shot transfer of the model across brain conditions (e.g., Alzheimer’s disease, schizophrenia, ADHD) and characterizes shared computational signatures.
Develops AI/ML methods for particle physics and the modeling of physical systems, including deep-learning and generative AI approaches to reconstruction, event classification, simulation and surrogate modeling, and analysis of large-scale detector and experimental data.
Adapts and transfers modeling approaches between the particle-physics and brain-health domains, contributing to the group’s cross-domain foundation-model methodology.
Contributes to open, reproducible research outputs, including the FAIR benchmark, versioned code releases, containerized pipelines, model cards, and documentation.
Trains and mentors graduate and undergraduate students in AI/ML methods, data pipelines, and reproducible-research practices.
Maintains technical and administrative support for the research project.
Analyzes and maintains data.
Conducts literature reviews.
Assists with preparation of reports, manuscripts, and other documents.
Installs, sets up, and performs experiments and computational workflows, interacting with students and other laboratory staff under the direction of the principal investigator.
Performs other related work as needed.