Machine Learning Engineer · Software Engineer
Toronto, Ontario, Canada
Machine learning engineer with a mathematics and statistics background, working on models applied to audio and on the evaluation needed to judge them. Alongside that I design and ship production software end to end — nine applications released to the App Store as sole engineer, all of them running entirely on-device.
On the App Store — 9
In beta
In development
Generates a valid, editable synthesizer preset from a natural-language description of a sound. Python, PyTorch: a domain transformer emits a typed intermediate representation, a mixture-of-experts predicts the 881 parameters, and a deterministic compiler guarantees every result loads in the target synthesizer.
Teaches Independent Component Analysis by having the listener separate mixed audio by hand while the underlying linear algebra updates alongside. Swift, SwiftUI.
Extracts perceptual descriptors — brightness, dominance, tonality — from a raw waveform. JavaScript with no dependencies: DC removal, Hann windowing, and a radix-2 DFT implemented against a written specification.
Locates reproducible linear structure in MNIST by decomposing digits into geometric fragments and tracing variance across them. Python, NumPy, scikit-learn, PCA and CCA.
Interactive dashboard mapping housing-need disparities across Toronto neighbourhoods. R, Shiny; reconciles census variables whose definitions changed between 2016 and 2021 so the two years can be compared at all.
Open-source protocol for coordinating multiple coding agents through executor and reviewer roles, with hand-off contracts that keep each step auditable.
Organises local sample libraries past the point a folder structure can handle. Python, SQLite, PyQt.
Honours Bachelor of Science, Mathematics & Its Applications Specialist (Probability / Statistics). Coursework in probability, statistics, optimization, linear algebra, and algorithm design. Requirements completed August 2026.
Awarded for an interactive Swift project on Independent Component Analysis.
50+ original tracks released and distributed globally, with support from Martin Garrix, Ferry Corsten, Anjunadeep Edition, and Monstercat Silk Showcase. Audio direction for four game titles, covering score, sound design, and implementation-ready assets. Cover art, logos, and application icons are also my own work.