University of British Columbia

Graduate Research Assistant · Sep 2017 – Sep 2019 · Vancouver, Canada

MSc in Biomedical Engineering — two years spanning hardware, software, and ML. Designed and built a physical cardiovascular benchtop system (servo pumps, sensors, LabVIEW real-time control with PID + feedforward). Built Python/TensorFlow ML models and pipelines to predict aneurysm rupture risk. Integrated CFD simulations (COMSOL, MATLAB) with machine learning for vascular treatment analysis.

Hardware: Cardiovascular Benchtop System

• Designed and built an in vitro cardiovascular benchtop system to replicate physiological abdominal aortic flow and pressure waveforms for aneurysm hemodynamics research.
• Architected a closed-loop flow circuit: servo gear pump, electromagnetic flowmeter, pressure transducers, compliance and resistance elements, and silicone AAA phantom — using commercially available components for low-cost, reconfigurable experiments.
• 3D-printed cores and plaster molds to fabricate silicone artery phantoms with controlled wall thickness.

Control Systems & Electrical Engineering

• Implemented a LabVIEW-based real-time control system using combined PID feedback and model-based feedforward control to track prescribed cardiac flow waveforms.
• Performed system identification in LabVIEW to derive a higher-order transfer function of the flow loop and synthesize an inverse-model feedforward controller to improve tracking performance.
• Achieved flow waveform tracking within ~2% error (R² > 0.998) and physiological pressure reproduction (80–120 mmHg) across resting and exercise cardiac profiles.

Machine Learning & Computational Modeling

• Developed Python/TensorFlow machine learning models to predict cerebral aneurysm rupture risk using hemodynamic and sensor-derived datasets.
• Built ML pipelines to analyze cardiovascular flow dynamics and evaluate stent treatment efficacy using experimental and simulated data.
• Integrated Computational Fluid Dynamics (COMSOL, MATLAB) with machine learning models to improve predictions of vascular treatment outcomes.
• Compared Newtonian vs. non-Newtonian blood flow models, highlighting shear-thinning importance for accurate rupture prediction.

Publications

• APS Division of Fluid Dynamics (2019) — 2D Particle Image Velocimetry and Computational Fluid Dynamics Study on Sidewall Brain Aneurysm.
• UBC MSc Thesis (2019) — Machine Learning-Driven In Vitro System for Cardiovascular Flow Analysis and Cerebral Aneurysm Hemodynamics Modeling.