Experience

UBC Data Science Institute

• Canada
• Research Services
• 1 - 100 Employee

• Data Scientist

  • Feb 2023 - Present

  • Developed a Machine Learning pipeline that automates free-text cancer report classification with near-human-level accuracy at 91% using BERT NLP for question answering and text classification. • Developed a Machine Learning pipeline that automates free-text cancer report classification with near-human-level accuracy at 91% using BERT NLP for question answering and text classification.



The University of British Columbia

• Canada
• Higher Education
• 700 & Above Employee

• Data Scientist

  • Sep 2020 - Dec 2022

  • Worked with Dr. Gabriela V. Cohen Freue and the Canadian Statistical Sciences Institute (CANSSI) as a data scientist to design and implement a simulation study comparing the prediction performance and variable selection performance of traditional survival analysis methods with the newer survival analysis method Casebase in various simulation settings, resulting in a new discovery showing penalized Casebase models to have up to 12% increased prediction performance compared to other survival analysis models.• My work further shows the performance advantages of Casebase, a novel survival analysis method that allows clinicians to measure the 5 or 10 year risk of disease diagnosis with better accuracy given a patient's covariate profile.• Presented my research Penalized Casebase in Survival Analysis to over 50 attendees, including the directors of the CANSSI Research Team and the directors of the UBC Department of Statistics. Show less

• Statistical Consultant

  • Jan 2021 - Apr 2021

  • Performed statistical analysis for a PhD candidate in Special Education research constructing mixed-effects models to compare the change in mental-distress over time across treatment and control groups.

• Teaching Assistant

  • Jan 2021 - Apr 2021

  • Taught statistics to university undergraduates.



Riversol

• Canada
• Wholesale

• Data Scientist

  • May 2020 - Aug 2020

  • Trained and validated an XGBoost classification model that was able to correctly predict 70% of purchase decisions out of all sample-takers that did purchase, 20% more accurate than previously trained models. • Addressed class imbalance in model selection and optimized the hyperparameters of 9 different classifiers. • Resulted in tailored marketing decisions for Riversol and acquisition over 25k new purchasing customers. • Trained and validated an XGBoost classification model that was able to correctly predict 70% of purchase decisions out of all sample-takers that did purchase, 20% more accurate than previously trained models. • Addressed class imbalance in model selection and optimized the hyperparameters of 9 different classifiers. • Resulted in tailored marketing decisions for Riversol and acquisition over 25k new purchasing customers.

• Research Data Science Student

  • Sep 2017 - Apr 2019

  • Worked with Dr. Frances Chen to investigate the association between visual attention towards social people and perceived feelings of connectedness. • Worked with Dr. Victoria Savalei to compare factor structure of alternative psychological scale formats and assess validity and reliability of each format using factor analysis and correlational analysis techniques in R. • Worked with Dr. Frances Chen to investigate the association between visual attention towards social people and perceived feelings of connectedness. • Worked with Dr. Victoria Savalei to compare factor structure of alternative psychological scale formats and assess validity and reliability of each format using factor analysis and correlational analysis techniques in R.