Designed and fine-tuned deep learning model parameters in PyTorch to improve registration accuracy. Created and validated algorithm in Matlab to remove intensity artifacts. Processed large whole slide images (>3 GB) using OpenCV and scikit-image for stain-agnostic registration. Conducted wet-lab experiments, in particular in vivo imaging with PARS technology and optical laser alignment.

Developed and improved two deep learning models with PyTorch to clean data and reduce error by 25%. Conducted data processing for two teams with pandas, NumPy, condensed timeline from 2+ weeks to 3 days. Extracted colour features from images in various colour spaces with OpenCV, with >75% label accuracy. Refactored and integrated preliminary research code to features in flagship application workflow.

Assisted with process modelling of lateral flow assay. Researched and gathered genes for glycolysis and oxidative metabolism in S. cerevisiae strains. Implemented and used yeast metabolic models.

Utilized ML models with various methods (SVM, KNN) to improve video signal feature-labeling and analysis. Extracted skin tone features from images in various colour spaces with OpenCV, using KMeans algorithm. Provided assistance with data cleaning and organizing tasks with automated scripts using pandas, NumPy. Detected outliers with statistical methods to improve peak-finding behaviour within flagship application.

Iterated soft robotics product for lymphedema using SolidWorks, AutoCAD, compressing timeline by 33%. Translated patent for entry to 2 international markets, enabled distribution plan with estimated $100k revenue. Generated novel PCB designs using Eagle, optimizing and consolidating testing pipeline with soldered parts. Presented data visualizations weekly to stakeholders, providing resources for cross-functional collaboration.

Worked alongside a team of 6 people to design mechanical arches for the UBC Mars Airlock Competition by organizing prior research and reaching out to suppliers for the production phase.

Assisting with infrastructure component porting for Android Automotive. Created two non-compliance dashboards using SQL and Jira macros for issue tracking and management, reducing active issues from >100 to 80 items across 3+ epics spanning several teams. Led development of Python service for signal simulation in Unreal Engine to reduce testing resources by ~20%.

Contributed to the Ford IVI system by focusing on the service-oriented architecture structure of the project. Developed and debugged Java, C/C++, Kotlin code under Agile design management. Created Android applications/services in Java and Kotlin to practice using JNI and Android NDK.