Health Data Scientist

University of Dundee; Dundee; On-site; GBP 50,000 - 75,000; Full time

Job summary

University of Dundee seeks two health data scientists to work with Pearson and Trucco groups on AI methods for EHR and imaging data to predict health outcomes in diabetes and study ageing effects. The role involves handling longitudinal data and developing predictive models.

You will collaborate across medicine and computing, publish results, and contribute to open-source code; prior experience with Python, ML, and healthcare data is essential.

Qualifications

• Excellent skills in AI and ML software development for data and image analysis.
• Experience with healthcare data and image analysis via machine learning.
• Experience in healthcare data curation and processing in safe havens.
• Strong English communication and writing abilities.
• Ability to wrangle EHR data and imaging data.

Responsibilities

• Wrangle electronic health record data and imaging data.
• Develop, optimize and validate models of biological ageing.
• Develop and validate prediction models for ageing-related outcomes.
• Ensure code is annotated and available in a public repository (e.g. GitHub).
• Write and present reports and presentations to stakeholders.

Skills

AI & ML

Python

R

SQL

Data visualization

Education

PhD in a numerate discipline

Tools

Git/GitHub

PySpark

TensorFlow

Job description

Full-Time for up to 24 months

Faculty of Health

We are seeking two exceptional health data scientists to join the groups of Pearson (Faculty of Health, Department of Medicine) and Trucco (Faculty of Science, Engineering and Business, Department of Computing) at the University of Dundee.

In this role you will develop and apply AI methods to electronic patient record data and linked retinal and other imaging data to develop, validate and apply models to predict health outcomes in people with diabetes. You’ll investigate the impact of accelerated ageing on these outcomes, and whether commonly increasingly used diabetes drugs like Ozempic, Mounjaro and Dapagliflozin reduce biological ageing. You’ll also work with linked molecular data – including genotyping array and proteomics.

About the Role

As a health data scientist you will join a team of clinicians with a long track record of working with population electronic health record (EHR) data (Pearson, Mordi) who work closely with colleagues from computing (Trucco, McKenna) with expertise in computer vision and AI image analysis. You will be expected to wrangle EHR data, and run frequentist and Bayesian statistical models, machine learning and deep learning models incorporating multimodal data. The strength, and challenge, of the data is the >25 years of longitudinal follow up and you will be expected to have or develop skills in longitudinal modelling, and development and validation of prediction models. The focus of this project is on biological ageing – with ageing clocks derived from retinal imaging, or proteomic analysis. We are interested in who has accelerated ageing, the clinical impact of accelerated ageing and whether routinely used drugs, like the GLP-1RA Ozempic, alter ageing rates. This work is a co llaboration with Eli Lilly.

Key Responsibilities

• Wrangling and curating electronic health record data and imaging data
• Developing, optimizing and validating models of biological ageing
• Developing and validating prediction models for ageing related outcomes
• Ensuring all code is fully annotated and available in a public repository e.g. github
• Writing and presenting reports and presentations to stakeholders

Essential

Excellent skills and demonstrable experience of AI and machine learning software development for data and image analysis, as well as use of current packages and languages e.g., Python and relevant libraries.

Demonstrable experience in software projects involving healthcare data and image analysis via machine learning.

Experience of healthcare data curation and processing in safe havens.

Presentation skills: demonstrable experience of successful paper writing, conference and workshop presentations.

Communication skills: excellent command of the English language and its use for successful, constructive interactions with other members of a research group as well as external collaborators, both academic and in companies.

Desirable

PhD in a numerate discipline

Knowledge and expertise in SQL, R, Python, PySpark

Knowledge and experience in analysing molecular data including proteomic and genomic data

Experience of working with Bayesian statistical methods

Active roles as a researcher in data science projects aimed to biomarker discovery

UK Visas and Immigration (UKVI) Requirements

Any offer of appointment will be conditional upon the individual having, or being able to obtain, the right to work in the UK. Prospective applicants are encouraged to review UK Government immigration guidance to assess their eligibility to work in the UK, including consideration of visa options that may be available to them, prior to applying.