POSTDOCTORAL RESEARCH FELLOW IN BIOINFORMATICS/MACHINE LEARNING

To conduct data harmonization across multiple data-types including electronic health records, digital pathology, clinical notes and omics measurements (bulk and spatial) for patients undergoing heart transplantation. This position is strongly focused on method and software development related to 1) learning representations from multimodal data and 2) developing predictive models of rejection trajectories using multimodal data. Major responsibilities: 1) benchmarking existing methodologies for time-series data imputation, clustering and classification, 2) developing methods and software for the analysis of partially observed, multimodal time-series data, and 3) writing research proposals, reports and publications.

• Performs data cleaning and analyses of time-series multimodal data such as clinical notes, laboratory values, imaging data and omics data often found in electronic health records (e.g. https://physionet.org/content/mimiciv/0.4/)
• Research and select appropriate analytical methods and software
• Develop analysis plans as needed
• Develop Python scripts for data cleaning, analysis, benchmarking and visualization
• Develops Python libraries leveraging existing libraries for the analysis of multimodal data (e.g. spacy, napari, squidpy, genism, torch, torch_geometric_temporal)
• Works on a team with other researchers, statisticians, and computational biologists in the identification and completion of assigned research projects
• Collaborates with other researchers, clinicians and statisticians to share knowledge and strategies for data analysis
• Provide computational expertise and advice as needed
• Contributes to the development of study design plans
• Develops, implements, tests and documents scripts and procedures to perform analyses
• Assists in preparation of conference presentations, academic manuscripts, grant applications, and technical reports, as required
• Contributes to training and supervision of undergraduate and graduate students

• PhD degree (received within the last 5 years) with proven interest in research
• Proficient in Python and implementation of different deep learning architectures (e.g. ANNs, CNNs, RNNs, GNNs) in PyTorch
• Effective oral and written communication and interpersonal skills
• Able to work independently within a team environment, work on multiple projects concurrently and work effectively under pressure to meet deadlines
• Prior experience with partially observed time-series data is an asset but not a must.
• Prior experience with high performance computing is an asset but not a must.
• Prior experience with omics data is an asset but not a must.