Methodology

How ML Signals Volatility

Machine learning volatility analysis session

ML Model-First Approach Each module centres on a specific model class — regression, ensemble, or neural — applied to real volatility data.

3 Study Tracks Available Modules span beginner to advanced, so participants start where their background actually is.

CA Remote, Canadian Context Curriculum examples draw from Canadian market data and regulatory framing, studied from home at your schedule.

Study Modules

Three structured modules, each targeting a distinct layer of machine learning applied to market volatility.

Machine Learning Intermediate

LSTM Networks for Short-Term Volatility Forecasting

A technical walkthrough of using Long Short-Term Memory networks to model time-series volatility patterns in equity markets, with attention to data preprocessing and model evaluation.

9 min

1,840 CAD One-time payment, lifetime access to materials

Price includes all session recordings, code repositories, and one round of instructor feedback per assignment.

11 seats remaining

Machine Learning Intermediate

Ensemble Methods for Volatility Regime Detection

How gradient boosting and random forest models can identify shifts between low-volatility and high-volatility market regimes using structured feature sets derived from price and macro data.

8 min

1,650 CAD One-time payment, access does not expire

Includes downloadable datasets, Jupyter notebooks, and two live Q&A sessions per cohort.

14 seats remaining

NLP & Market Analysis Advanced

NLP-Based Sentiment Signals in Volatility Analysis

Applying transformer-based NLP models to financial news and earnings call transcripts to extract sentiment signals that correlate with near-term implied volatility movements.

10 min

1,980 CAD One-time payment, cohort-based with fixed start dates

Price covers course access, data licenses for training exercises, and one-on-one project review session with the instructor.

8 seats remaining

How Each Module Is Structured

Modules follow a consistent format so participants spend less time orienting and more time working with data. Each session builds on the previous one without gaps.

All exercises use publicly available Canadian market datasets. No proprietary platforms or paid subscriptions are required beyond the module fee.

01

Concept framing

A short lecture segment establishes what the model does and where it breaks down — both matter equally.
02

Dataset walkthrough

Participants examine a real volatility dataset, noting its quirks — missing values, regime shifts, outlier sessions.
03

Model implementation

Step-by-step coding in Python. Each line is explained; no prior ML library experience assumed at the entry level.
04

Diagnostic review

Participants interpret output metrics — residuals, feature importance, validation curves — and identify where the model is unreliable.

VIX Volatility Index

Volatility is not random noise — it has structure ML can detect

Market volatility clusters in time, responds to macro signals, and exhibits memory effects. These patterns are detectable with the right model and enough clean data. That is the starting premise of every module offered here.

Who Teaches These Modules

Mireille Fontaine

Quantitative Methods Instructor

Has spent several years applying gradient-boosted models to volatility forecasting for fixed-income portfolios. Teaches the ensemble methods track.

Aleksandra Wiśniewska

Data Science Lead

Background in statistical signal processing; leads the regression fundamentals and neural network modules. Works primarily with time-series financial data.

Inside a Working Session

Participants working through a data exercise

Sessions run asynchronously

Recorded lectures and exercises are available on demand. Participants in Gatineau and across Quebec complete modules at their own pace, replaying segments as needed — no fixed schedule required.

Code walkthrough during a volatility module