Research Methodologist, Data Science

The Research Methodologist for Data Science will be a key member of the research team in the Regulatory Science Lab at the University of British Columbia’s (UBC) Academy of Translational Medicine (ATM). They will contribute to ongoing and emerging research activities augmenting health data and data systems to enable rapid evidence generation for innovative health products. The Research Methodologist will support data preparation and application of natural language processing (NLP) and other machine learning (ML) methods, including training and deploying large language models (LLMs) for semi-structured electronic health records (EHRs).

They will also undertake benchmark implementation and statistical analysis for performance evaluation. Other responsibilities include maintaining reproducible code repositories, supporting grant applications, writing reports and manuscripts, presenting results to multidisciplinary audiences and other forms of knowledge exchange. The working environment will require both independent research and working as a core member of project teams, with opportunities to engage with interdisciplinary researchers, health care stakeholders, and policy makers.

This is a full time 1-year term position, with the possibility of extension.

Organizational Status

The Research Methodologist for Data Science works independently and reports directly to Dr. Dean Regier on technical and project management aspects. The incumbent will interact and work with other faculty, research staff, and graduate students within the Regulatory Science Lab and the ATM, and external collaborators.

Housed within the Faculty of Medicine, the School of Population and Public Health (SPPH) is an innovative unit that encompasses many of the health-related groupings at UBC as a collaborative venture. The School is structured around four divisions:
Occupational and Environmental Health;
Health Services and Policy;
Epidemiology, Biostatistics and Public Health Practice; and Health in Populations. The resulting mix of professions and disciplines is seen as a means of connecting individuals and learners to galvanize the relationship between health research, public health and health services and to enhance learning.

The ATM at the University of British Columbia (UBC) is established by the Faculty of Medicine. The ATM is a nucleus for translational and regulatory science research, education and training, and innovation. The ATM is a strategic imperative for the Faculty of Medicine and has a mandate to collaborate with other units in the Faculty of Medicine, the wider University, and beyond to accelerate the translational medicine continuum.

The ATM will drive impactful medical and policy research to create new knowledge to improve health outcomes and benefit society. The ATM will be a nucleus for the development of new academic and training programs, creating a cutting-edge ecosystem in which top educators and researchers train the next generation of health innovators. The ATM has supported the recruitment and establishment of over 36 new faculty.

Work Performed

• Develop, program, test, validate, and report on ML pipelines, including LLMs, and other machine learning models for heterogeneous health data extraction and augmentation to support downstream research questions.

• Develop and implement benchmarks and continual model performance monitoring systems across various data sources and scenarios while assessing safety and risks.

• Review internal databases and code repositories and ensure compliance with ethics, privacy, and data security requirements.

• Conduct bias audits to ensure model predictions are fair across all relevant data sources.

• Stay up-to-date with reviewing and potentially integrating new approaches and techniques from emerging scientific literature on language models, optimization, and evaluation techniques.

• Define the scope and data elements required for machine learning projects.

• Document data and methodologies of projects, including on data sources, training protocols, and compliance with data safety and ethics requirements, and ensure project deliverables are met.

• Develop scripts and data processing pipelines using R or Python, leveraging advanced expertise for data processing, model development, and data visualization for both model training and deployment.

• Work with senior staff members to develop and modify models, and mentor junior staff and students, as needed.

• Develop, write, coordinate, and edit scientific manuscripts, reports and other knowledge translation products, including design, compilation, synthesis, dissemination and evaluation.

• Contribute to development of new research proposals and other knowledge translation activities.

Consequence of Error/Judgement

The incumbent is given wide latitude for exercising independent initiative and judgment in performing specialized duties and responsibilities. A lack of judgment could harm the research team and partner organizations' research and funding. The incumbent…