Scientific Leadership Course, 2008 – University of Calgary, Calgary, Alberta
Ph.D. in Medical Biophysics, 1996 – Western University, London, Ontario (Collip Medal: top PhD grad)
Master of Science in Computer Science, 1989 – University of Regina & The Allan Blair Cancer Center, Regina, Saskatchewan
Bachelor of Science Honors in Computer Science, 1986 – University of Regina, Regina, Saskatchewan (a Co-op Work Study program. Graduated with High Honors: GPA > 85%)
Communication: Accomplished & engaging public speaker. Able to effectively convey complex technical concepts to non-expert audiences.
Programming: Python, C
Medical Imaging: Teleradiology, MRI, CT, PET, US, Mammography, DICOM, Radiomics, Biomarkers
Applications: Cancer (brain, ENT, breast, liver, pancreas) MS, Stroke, Alzheimer’s, Neurosurgery
Image Processing: Computer Vision, ITK, VTK, OpenCV, scikit-image, Wavelets, Fourier, Stockwell, Statistics, Analytics, Algorithms, Graphics, GPGPU
ML: PyTorch, Keras, TensorFlow, huggingface.co, Spacy, scikit-learn, XGB, Random Forests, PCA
Technology: AWS, Docker, linux, bash, MacOS, iOS, Xcode, Windows, make, git, latex
H. LEE MOFFITT CANCER CENTER AND RESEARCH INSTITUTE, Tampa, Florida
(Inaugural) AI Officer and Senior Member of Research Staff, January 2019 to May 2021
Lead Center efforts to develop digital tools that utilize artificial intelligence and other advanced technologies to improve the efficiency and quality of cancer care. Founding Chair of the Clinical AI Committee that included clinicians from across the practice. We identified novel opportunities with high impact, developed innovative solutions, and oversaw translation to clinical deployment.
Developed multiple models to predict patient outcomes using EHR data. Applications include predicting sepsis onset in bone-marrow transplant and malignant hematology patients, 60 - 180 day survival in lung cancer patients, and more. Models use recent techniques to accommodate the sparsity and irregularity of EHR data. These include: GANs, VAEs, SplineCNNs. Applied methods to explain model performance: SHAP, feature ranking. Built dashboards to communicate results to healthcare providers: Dash, Plotly, Streamlit.
Developed a novel recurrent neural network architecture (feature + time attention → BiLSTM → multi-head decoder) to predict sepsis onset in ICU patients. Have achieved 95% accuracy, surpassing the previous state-of-the-art on the same widely used dataset (MIMIC-III) by 10%.
MAYO CLINIC COLLEGE OF MEDICINE, Phoenix, Arizona
Professor of Radiology, September 2011 to January 2019
Developed innovative and effective machine learning methods for precision health. Identified, extracted, combined, catalogued, and distributed information from medical images and signals. Actively pursued research through collaboration with geneticists, pathologists, and epidemiologists. Fostered the creation of sophisticated tools based on visualization and machine learning to help locate and identify patterns that correlate with clinical outcomes. Enhanced insight into origins of disease, aided diagnosis, and improved delivery of treatment.
Saved Mayo Clinic $10M on Mobile Image Viewer by developing a research partnership with PureWeb Inc., vendor of Mayo Clinic’s enterprise mobile imaging solution.
Guided project to develop informatics tools for Division of Breast Imaging resulting in turnaround time on BI reports dropping from 24 hours to 27 minutes.
Collaborated with advisory panel to submit a document to senior Mayo leadership proposing a new machine learning and data science center.
ALBERTA INGENUITY CENTER FOR MACHINE LEARNING, Edmonton, Alberta, Canada
Fellow, July 2008 to September 2011
Early scientist in the field of radiomics (machine learning to discover medical imaging biomarkers). Focused research on medical imaging informatics while devising new methods to link imaging and non-imaging data to aid healthcare professionals and their patients. Investigated effective ways of understanding, diagnosing, treating, and monitoring disease while managing and coordinating healthcare with reduced errors.
PUREWEB, INC., Calgary, Alberta, Canada
Co-Founder and Founding Scientist, January 2004 to September 2011
Translated research into positive patient outcomes by serving as Co-Founder and Founding Scientist of company that develops ResolutionMD, an FDA Class II cleared product for state-of-the-art mobile tele-radiology. Supervised tasks related to cultivating the prototype, intellectual property, product management, sales and marketing, and regulatory clearance. Products are available in 13 languages and are installed in >2,500 healthcare facilities across 40 countries.
ResolutionMD currently used by the US Veterans Affairs (100k employees), Intermountain Health Utah, Nebraska Medicine, Mozi Healthcare, Massachusetts General Hospital, Phoenix Children’s Hospital, Mayo Clinic, and others, for their enterprise image viewing.
Market capitalization in excess of $150M.
UNIVERSITY OF CALGARY, Calgary, Alberta Canada
Professor of Biomedical Engineering, Radiology and Clinical Neurosciences, July 2000 to September 2011
WESTERN UNIVERSITY, London, Ontario, Canada
Assistant Professor, Diagnostic Radiology & Medical Biophysics, October 1996 to June 2000
23 Scientists and Technical Staff
12 Post-doctoral Fellows, Medical Fellows and Residents
31 Graduate Students (15 PhD and 16 MSc Students)