About
Applied Mathematics • Time Series • Machine Learning
I'm a student in Mathematics and Computer Science at Université Laval and a research intern at INRS, where I work on modeling and forecasting environmental time series.
What drives me: understanding models deeply — their assumptions, their real behavior, their limits — and translating them faithfully into concrete, verifiable implementations.
What I focus on
Time series
Forecasting multi-site sensor data: SARIMAX, long-memory models, diagnostics, and rigorous out-of-sample validation.
Machine learning
Neural networks for time series (TCN), hyperparameter search on compute servers, and honest comparison against statistical baselines.
Applied mathematics
Continuous and combinatorial optimization, numerical linear algebra, ODE/PDE solvers, algorithm analysis.
Scientific visualization
Bilingual interactive articles with Manim-style animations: copulas, wavelets, GARCH, reinforcement learning.
My approach
Baseline first. A complex model earns its place only if it beats a simple, well-specified, properly validated one. I apply this discipline in research and in my personal projects alike.
A significant part of my learning is self-directed. I go back to definitions and core assumptions, then document what I build — code, LaTeX reports, interactive articles.
Current work
Research (INRS). Modeling environmental time series from multi-site hourly sensor data: statistical baselines (SARIMAX, long-memory models) benchmarked against deep learning architectures on high-performance computing infrastructure.
Interactive blog. A bilingual series of articles that make advanced mathematical concepts visible through interactive animations, including a live copula exploration laboratory.
Read the interactive articlesCurrently
Studying
- Optimization (continuous and combinatorial)
- Data analysis
- Numerical solution of ODEs and PDEs
Building
- Forecasting pipelines (INRS research)
- Bilingual interactive articles (Manim style)
- Copula laboratory and visualization tools
Goals
- Take models from research to production (deployment, monitoring)
- Master modern time series architectures (attention, transformers)
- Publish rigorous interactive articles regularly
Contact
Open to collaborations, technical discussions, and projects related to time series, machine learning, and applied mathematics.