Student Spotlight: Jillian Melbourne, MS in Health Data Science ’26
Jillian Melbourne was first drawn to health data science while researching patient-specific algorithms for deep brain stimulation for Parkinson’s disease in Helen Bronte-Stewart’s lab at Stanford University, where she saw how data-driven methods could be tailored to an individual patient. Adaptive deep brain stimulation, which functions like a brain pacemaker, adjusts stimulation in response to signals recorded from a patient’s brain and body. Seeing what a well-tuned algorithm could accomplish for one patient made Melbourne wonder about how similar approaches could reach patients beyond specialized medical centers.
“Most people with Parkinson’s will never see a movement disorders specialist, let alone have access to personalized deep brain stimulation,” Melbourne says. “That question pushed me toward research built on widely available health data, where methods can generalize across care settings rather than depending on specialty care.”
That interest has continued to shape Melbourne’s work. Before coming to Dartmouth, she worked on deep learning frameworks for medical image segmentation and developed models that predict menstrual cycle patterns from wearable biometrics in diverse populations. These experiences strengthened her interest in developing methods that can extract meaningful insights from health data collected in real-world settings.
“I came to Dartmouth to develop the technical skills necessary to apply statistical methods to messy, complicated patient data,” she says.
The Online MS in Health Data Science program at Geisel, with its combination of classical biostatistical methods and machine learning, was particularly appealing to Melbourne. She was drawn to the opportunity to work with institutional clinical data through Dartmouth Health, an experience that ultimately led to her receiving the Samuel D. Pressman Fellowship.
The fellowship is a research award through Dartmouth Health’s Department of Radiology, the Geisel School of Medicine, and the Center for Precision Health and AI. Melbourne’s project focuses on developing a multimodal framework that combines deep learning features from radiology (MRI) and pathology (microscopic imaging of tissue biopsies). The premise is that the two imaging modalities capture complementary aspects of a patient’s biological information. By combining deep learning features from both, Melbourne hopes to develop models that can better predict the risk of breast cancer recurrence.
To validate the models, Melbourne plans to compare radiology-only, pathology-only, and multimodal approaches against recurrence events identified through clinical chart review, led by Geisel medical student Ian Beals. The project brings together expertise from across Dartmouth, with mentorship from Roberta diFlorio-Alexander in radiology, Jonathan Marotti in pathology, Brittny Calsbeek in Biomedical Data Science, and Sarah Preum in Computer Science.
Before joining the Health Data Science cohort at Dartmouth, Melbourne was involved in clinical trials and data collection. That experience gave her a close understanding of the research process, but it also revealed a gap in her own training.
“I lacked some of the intuition for how to best analyze the data I was collecting,” she says.
At Dartmouth, her training in statistical and computational methods has helped bridge that gap. “Dartmouth has given me the formal methods training to make those methodological decisions myself and to build pipelines from existing clinical data. That has shifted what I want my career to look like, from managing research to developing the methods behind it.”
Reflecting on her student experience Melbourne says, “I expected Dartmouth to attract accomplished people, but I have still been struck by just how broad my cohort is.” Her classmates include practicing neurosurgeons, software engineers, data scientists, and more. Each brings a different perspective to questions about health data, and Melbourne has found that those perspectives can fundamentally change how a problem is approached.
“The same problem or dataset is approached completely differently by a clinician and an engineer, and I have learned as much from those exchanges as I have from the coursework itself,” she says. “I am grateful to learn with and from such an incredible and thoughtful group of people.”
Looking ahead, Melbourne plans to continue working on methods that use routinely and remotely collected clinical data to predict health outcomes. Her long-term goal is to develop tools that can work across diverse healthcare settings. Whether that work ultimately takes place in academia or industry is less important to her than remaining close to the methodological questions themselves, and she expects the next few years to involve additional training as she continues developing the skills needed to tackle those problems.
Her advice to students entering health data science reflects the same willingness to pursue difficult questions that has shaped her own path.
“I would tell incoming students that your past experience matters far less than pursuing work that excites you. Graduate school is the time to chase the questions you find most compelling, regardless of how good you are at them initially. I would encourage new students to be fearless when deciding what kind of research or output they want to produce, even if the learning curve is steep, and to reach for opportunities before you feel totally qualified for them.”
Written By: Mia Soucy
POSTED 9/28/2026 AT 03:17 PM IN #studentspotlight #HealthDataScience
GET IN TOUCH
To arrange a media interview, please contact:
geisel.communications
@dartmouth.edu