Daniel's Portfolio

Hey, I'm Daniel.

I am a machine learning researcher focused on developing and applying state-of-the-art techniques to address real-world challenges.

Check out my projects

About Me

I am a machine learning researcher based in Vancouver, British Columbia, who will soon have a Master's degree in Mathematics and Statistics from McGill University (conferral in May 2026). I also hold an Honours Bachelor of Science in Data Science from the University of British Columbia. My research focuses on Bayesian statistics and statistical machine learning, and I have worked in several applied domains including medical imaging/physics, curriculum design, natural language processing, and anomaly detection. You can browse my projects to see examples of this work.

I am currently open to both full-time and freelance opportunities. I am interested in solving challenging and impactful real-world problems. If you have something in mind, please feel free to contact me. Want to see my CV? Click here to see it.

My Projects

Fuzzy Medical Image Segmentation

This project showcases a medical image segmentation system that combines fuzzy clustering with centroid initialization techniques to separate structures, such as tumors, in mammography data. Using soft clustering allows us to quantify the uncertainty in pixel assignments, providing more reliable and interpretable segmentations.

Spectral Biochemical Identification

Raman spectra act as a biochemical fingerprint comprised of overlapping component signals. This project separates those blended signatures using a Bayesian approach to matrix factorization. This produces clear component spectra along with uncertainty estimates for reliable chemical identification.

Functional Anomaly Detection

This project develops a Bayesian nonparametric model that treats time-series as functions and identifies unusual behaviour using infinite mixtures of multi-output Gaussian processes. Wavelet-based smoothing and automatic kernel selection allow the system to detect subtle anomalies with calibrated uncertainty.

CurricularAnalytics

An R package implementation of the Curricular Analytics framework for analyzing and visualizing curriculum structures. The package provides tools to quantify structural complexity through metrics like blocking factor, delay factor, and centrality, helping educators optimize curriculum design and identify potential bottlenecks in degree pathways.