David Landsman – Computer Programmer

David Landsman (Computer Programmer) LinkedIn ORCIDGitHub
David is a recent graduate from the Computer Science program at the University of Toronto. He loves studying the theory behind various fields in Computer Science, including Machine Learning and Cryptography, as well as applying his knowledge to various real-world problems. David started working at the MAP Centre for Urban Health Solutions as a computer programmer via the University of Toronto work-study program with supervision from Sharmistha during his undergrad studies. He developed and implemented an algorithm for assigning clinicians to on-call service for the Infectious Diseases department at St. Michael’s Hospital. He collaborated with the head of the Infectious Disease division in developing and testing the algorithm. Since 2018, David’s application has been used to generate the clinical schedules in the department His tool is publically available alongside the manual for use, and his paper from the work is currently in preprint:

  • Landsman D, Ma H, Knight J, Gough K, Mishra S. (2019). A flexible integer linear programming formulation for scheduling clinician on-call service in hospitals. [Preprint available].

David then joined the Mishra Lab as a part-time computer programmer and research staff in October 2019 to lead the development of a retrospective and prospective clinical cohort using structured and unstructured data from clinical and electronic health records. Specifically, David leads the Tuberculosis Database project in collaboration with Dr. Jane Batt (clinical director of the St. Michael’s Hospital Tuberculosis Clinic). The project is being conducted in collaboration with LKS-CHART . The database collects retrospective patient data from electronic health records, including medical history, test results, diagnosis and medications. A novel component of the database involves extracting useful variables from unstructured clinician dictations. David is exploring a variety of machine learning approaches to tackle this problem, including rule-based natural language processing and unsupervised clustering. The Tuberculosis Database will be a resource for scientists conducting research into clinical epidemiology, quality improvement and implementation science research.