Project Portfolio
A collection of software engineering and data science projects showcasing full-stack development, AI/ML applications, and statistical analysis.
AI Model Evaluation and Developer Tooling at IBM watsonx
Working on AI model evaluation and internal full-stack tools for enterprise coding assistants. Evaluated 100+ LLMs, built React dashboards, and improved evaluation workflows through automation and LLM-assisted tooling.
Fine-Grained Phytolith Image Classification for Wheat Species
Developing computer vision models to classify multi-cell phytolith images across different wheat species. Built classification baselines using VGG19, ResNet-18, Vision Transformers, and YOLOv8 with reproducible training workflows.
UTBook: Computer Science Learning and Assessment Platform
Full-stack online learning and course-management platform for CS education. Features assignment submission, automated testing, interview coordination, and office-hour booking in one centralized system.
UTM Teaching Assistant Application Platform
Web platform designed to improve the TA application workflow at UTM. Centralizes applicant information, course preferences, and administrative review workflows in one system.
Online Auction Platform Development
Contributed to development and maintenance of an online auction platform. Developed responsive frontend features using React and Material UI, integrated with Django REST Framework APIs.
Animal Rescue Website Development
Volunteer software engineer for Save Fur Pets Animal Rescue. Developed and maintained website features, created Cypress end-to-end tests, and supported animal rescue operations through technical work.
Interactive Programming Challenges for CSC309
Designed hands-on programming challenges and automated testing workflows for web-development tutorials. Created unit tests and automated feedback using GitHub Classroom and Vitest for ~60 students per tutorial.
Food Preference Classification Using Machine Learning
Machine learning project predicting food preferences based on survey responses. Compared multiple classification models and achieved ~84% accuracy on hold-out test set using neural networks.
The Interplay Between CPI Changes and Vehicle Theft Frequencies
This study employs multiple regression models to identify correlations between various CPI categories and the frequency of vehicle thefts in Toronto, offering insights that could influence economic and crime prevention policies.
Forecasting the American Voters' Choices in 2018
This study examines the predictive power of income and gender on voting behavior in the 2018 US elections using the Cooperative Congressional Election Study (CCES) 2018 dataset.
NFL Passing Efficiency Forecasting
Analysis and predictive modeling for the 2023 NFL season, focusing on forecasting the passing efficiency of quarterbacks using the passing expected points added (passing_epa) metric.
Modeling the Efficacy of Possession in Football
This study examines the impact of ball possession on scoring outcomes in Bundesliga matches (2023-2024 season), employing both Poisson and negative binomial regression models to unravel the relationship between home team ball touches and goal scoring success.
Navigating Ethics and Engagement in Visualizing the Holocaust
An interactive Shiny app providing visualization of Holocaust victims' ethnic origins and numbers, specifically focusing on those murdered at Auschwitz concentration camp. This project aims to honor victims' memory through dignified data representation.
Interested in Working Together?
I'm always open to discussing new projects, creative ideas, or opportunities to be part of your vision.