Maxime Heuillet

Ph.D. Candidate in Computer Science, Université Laval & MILA

GitHub Google Scholar LinkedIn Download CV

About Me

I am a Ph.D. candidate in Computer Science at Université Laval and MILA – Québec AI Institute in Montréal (Canada), advised by Audrey Durand. My research lies at the intersection of interactive learning and statistical learning, with emphasis on theoretical foundations and practical applications. I am particularly interested in building efficient, robust and generalizable ML systems. I interned at the Thales (CortAIx lab), Noah's Ark lab, and Amazon Web Services in Seattle. I hold a M.Sc. from the National School of Statistics and Data Analysis (ENSAI) in France. Contact: maxime[.]heuillet1[at]ulaval.ca

Research Interests

• Interactive learning (bandits, partial monitoring, active learning, gflownets)
• Reinforcement Learning
• Robustness
• Efficiency
• Large Language Models
• Recommender Systems
• Automated Machine Learning

Selected Publications

• Robust Fine-Tuning from Non-Robust Pretrained Models: Mitigating Suboptimal Transfer With Epsilon-Scheduling, ICLR Main Track, 2026.
  Jonas Ngnawe, Maxime Heuillet, Sabyasachi Sahoo, Yann Pequignot, Ola Ahmad, Audrey Durand, Frederic Precioso, Christian Gagné
• Randomized Confidence Bounds for Stochastic Partial Monitoring, ICML Main Track, 2024.
  Maxime Heuillet, Ola Ahmad and Audrey Durand.
• Neural Active Learning Meets the Partial Monitoring Framework, UAI Main Track, 2024.
  Maxime Heuillet, Ola Ahmad, Audrey Durand.
• Workshop publications at NeurIPS, ICML, AAAI, etc.