Hi, I'm Ryan Divigalpitiya

Since June, 2025 I've been leading the development of an intelligent data pipeline and optimization platform for Calumix's graphene coating machines — integrating real-time sensor ingestion, automated data validation, statistical process control, and advanced machine learning techniques.

For my Master's directed study, I focused my research on Transformers, where I developed interpretability techniques to assess the ability of Transformers to capture the meaning of words from highly unique vocabularies — examples include slang and jargon found on social media, and industry-specific terms found in various industries.

In my spare time, I’m developing a weight-lifting workout app that will use a locally-deployed Large Language Model that will generate workout plans for specific muscle groups and convert routines written down in the iOS Notes app into workout plans managed within the app.

For my second startup, I built a convolutional neural network paired with thermal sensors that were then used to detect room occupancy in a smart-home. My early prototype achieved 99% accuracy in discriminating between false-positive and true-positive signals.

I co-founded a company that set out to build highly energy-efficient smart homes. Our builds utilize advanced technology to deliver more affordable, environmentally-sustainable living. The company is growing and expanding to build in various locations across Ontario. I left in 2018 to start Nexus Labs.

For my CS 9860 Advanced Machine Learning Course, I proposed and built Convolutional Neural Networks paired with inexpensive thermal sensors to recognize room occupancy in a smart-home. My model achieved ~99% accuracy in discriminating between false-positive and true-positive signals.

In collaboration with Fitbit©, my final project in CS 2212 Introduction to Software Engineering was to build a Fitbit© desktop application for Windows that could link to a user’s Fitbit© device. I was project manager and my team won 1st place for the best app in our class.

For my CS 9637 Data Science Course, I proposed, built and tuned a random-forest multi-classifier that can predict the probability of someone falling into a class of obesity given that they maintain their current lifestyle choices and habits. My model achieved 87% accuracy in predicting obesity.

In collaboration with the Schulich Medical School and the Department of Computer Science at Western University, my final project in CS 3307 was to build the next version of a data visualization application used by the medical school at Western. I was project manager for my team and developed back-end features.

For my CS 9552 Human-Computer Interaction course, I was the project manager for my team's final project. I proposed and took lead on designing an app that could help students identify careers that interest them and help them plan their academic path to best prepare for those careers.

On the side, I took several freelancing jobs in software development, UI / UX design, and graphic design. A startup called Normap (Toronto, Ontario) hired me as a freelancer to create UI designs of their future dashboard web app.

A company called Cylix (London, England) hired me as a freelancer to (1) create a refreshed UI design for their aging desktop and mobile interface and (2) develop their new front-end of their desktop app using Qt (in C++) based on my UI designs.