Geoff Klein

Experience

• 

  Royal Military College of Canada/Collège militaire royal du Canada

  May 2015 - Aug 2015

  Summer Research Student; Finite Element Modelling for Applied Electromagnetics

  • Developed Finite Element Method models for an electromagnetic probe used in the CANDU nuclear reactors.• Finite Element Method models invoked geometric approximations to allow for rotational symmetric and simplifications, achieving an error of 6% compared to experimental measurements.

• 

  Queen's University

  May 2016 - Aug 2017

  Graduate Student Researcher; Analytical Modelling for Applied Electromagnetics

  • Created an advanced mathematical model in MATLAB, to access a measurement tool used in the CANDU nuclear reactors. The probe and model utilized principles of low-frequency electromagnetic induction. The model was the most accurate available and allowed further quantification and development of the measurement tool.• Validated mathematical model through experimental measurements and a Finite Element Method also created, achieving only a 2% error between experimental and analytical model measurements.• Constructed a Finite Element Method model of the same measurement probe without any geometric approximations, achieving a 1% error evaluated against experimental measurements.• Assessed electromagnetic properties of materials in the CANDU nuclear reactors through experimental design.• Articulated complex research outcomes to diverse audiences at national conferences and regulatory bodies, such as the Canadian Nuclear Safety Commission, to highlight the progression of project milestones. Show less

• 

  Sunnybrook

  Sept 2017 - Jul 2024

  Computer Vision PhD Researcher; Medical Imaging and Deep Learning Researcher

  • Designed convolutional neural network models in Python to quantify and analyze clinical outcomes for vertebral stability for patients with spinal metastases.• Developed state-of-the-art vertebral detection, segmentation and clinical classification models using patient 3D CT scans.• Maximized model efficiency and reduced memory consumption by 50%, enabling model deployment in seconds on consumer-level computer hardware.• Partnered with key stakeholders to handle highly sensitive patient data; engaged 10+ specialists to refine clinical models, achieving 87% accuracy and ensuring compliance with clinical use-cases.• Collaborated with 10+ researchers and medical professionals to identify clinical requirements, aligning project goals.• Validated models with 6+ radiation oncologists and spine surgeons, iteratively updating based on feedback, ensuring fit-for-purpose solutions.• Orchestrated aligning objectives among co-op /graduate students and research engineers.• Communicated groundbreaking research to professionals at international conferences such as the Computational Methods and Clinical Applications for Spinal Imaging (Sept. 2019) and International Society of Orthopaedic Centers (Nov. 2022), enhancing global research networks. Show less