§05

ABOUT

Background

I'm a Montréal-based data scientist. I started in neuroscience, moved into engineering, and now build data systems end to end — collection through modelling through the interface someone actually uses.

At Dr. Jerbi's lab I built decoding pipelines for MEG neuroimaging data, extracting millisecond-resolution information about neural activity from raw sensor signals. That work shaped how I approach feature engineering, signal processing, and translating domain knowledge into ML pipelines.

At Artelys I applied the same rigour in an industrial setting: gradient-boosted models over technical transformer specifications to predict labour times, with a feature-importance layer so non-technical stakeholders could interrogate the output rather than take it on faith.

Since June 2026 I've been the sole data scientist on a lead-generation portfolio at Valnet, responsible for the whole lifecycle rather than a slice of it. The part I care most about is verification: refusing numbers that can't be reconciled, and treating a plausible result as a hypothesis until its noise floor is known.

Skills

Machine Learning & Analysis
XGBoostscikit-learnPyTorchTensorFlowHuggingFaceCausal inferenceExperimental designEmbeddings & clusteringLLM evaluationMNE-Python
Data & Cloud
BigQueryGoogle Cloud PlatformCloud RunSQLPostgreSQLDimensional modellingFirebaseMongoDB
Languages
PythonSQLTypeScriptRJavaMatlabC++
Web
ReactNext.jsNode.jsAngular