Zacharie Garnier-Cuchet

Hi! I’m Zacharie Garnier-Cuchet, an AI researcher and M.Sc. student in computer engineering at Polytechnique Montréal and MILA - Québec AI Institute, supervised by Amal Zouaq and Prof. Philippe Langlais.

I began in French engineering schools studying mathematics and physics, but discovering NLP transformed my path entirely. When large language models first emerged, I was captivated by machines capable of coherent reasoning, a moment that reshaped my academic identity and drove me to understand how intelligent systems learn and reason.

My research focuses on:

• Legal NLP — Automatic detection and resolution of legal references in judicial decisions, involving structured reasoning, document parsing, and representation learning for complex texts.
• Representation & Contrastive Learning — Building embeddings and semantic models for large-scale text understanding.
• Learning Dynamics — Contributing to optimization research, including a NeurIPS workshop paper on loss landscapes and network stability.

I also serve as a teaching assistant for NLP at Polytechnique Montréal.

Growing up between Montréal and rural France taught me how environment shapes thought and values. This awareness now drives my interest in Trustworthy AI, interpretability, and alignment, building transparent models that support democratic values rather than amplify inherited biases.

How do models acquire internal structure? How can learning systems amplify human insight over bias? How do we build AI for a safer, more equitable society?

If these questions resonate with you, I’d love to connect. Find my publications and contact information below.