Machine learning meets market analysis
How Domain came to exist
Domain was established in 2021 in Gatineau, QC, by a group of researchers who found that existing educational resources on algorithmic market analysis were either too abstract or too shallow for practitioners.
The focus settled on a narrow but demanding problem: applying machine learning to volatility modelling — specifically how ensemble methods, recurrent architectures, and feature engineering interact when price behaviour becomes non-stationary.
Lectures are structured for remote learners. Each session builds directly on the previous one, so the conceptual load accumulates gradually rather than arriving all at once.
Volatility Modelling
GARCH variants, realised volatility estimators, and regime detection form the analytical core of every module.
ML Architectures
LSTM, gradient boosting, and attention-based models are examined in the context of time-series with fat tails and structural breaks.
Remote Format
All sessions are delivered online, designed for learners who need a structured schedule that fits around existing professional commitments.
The people behind the lectures
Two instructors with complementary backgrounds — one in quantitative finance, one in applied machine learning — developed and deliver the full programme.
Étienne Marcellin
Quantitative AnalystSpent eight years at a fixed-income desk before shifting to research. His lectures focus on volatility term structure, GARCH specification, and how model assumptions break during stress periods.
Sofía Wrzesień
ML ResearcherHer background is in statistical learning applied to high-frequency data. She covers feature construction, model evaluation under distribution shift, and practical deployment constraints that textbooks tend to omit.
Programme in numbers
Delivered across 12 modules, each closing with a working code example in Python. Sessions are recorded and remain accessible throughout the enrolment period.