Head of the Fundamental AI Research group. Leading 20+ researchers and 30+ interns on multimodal learning, web agents, reasoning, and data analytics. Key projects include BrowserGym, WorkArena, BigDocs, and the Apriel open models.

Led the Low Data Learning team following the Element AI acquisition. Directed 5 researchers and 15 interns on self-supervised, semi-supervised, and weakly-supervised learning.

Research on segmentation, object counting, image generation, domain adaptation, and continual learning. Promoted to Team Lead in 2020; team transitioned to ServiceNow after acquisition.

Research on Indigenous language translation, supervising doctoral students, and teaching activities.

Affiliated with MILA through Polytechnique Montréal, contributing to collaborative research in multimodal learning and AI agents.

Instructor in the Computer Vision Master’s program, teaching graduate courses and supervising Master’s theses.

Deep learning for image generation and semantic segmentation under Prof. Aaron Courville and Prof. Antonio M. López. Awarded the Marie Curie European Fellowship.

Led the Elektra autonomous vehicle project and developed the SYNTHIA autonomous driving simulator. Partnerships with Tazzari and NVIDIA.

Domain Adaptation of Virtual and Real Worlds for Pedestrian Detection. Advisor: Prof. Antonio M. López. Recipient of four best thesis awards (IEEE ITS Society, AERFAI, UAB, PIONEER/CERCA).

Thesis on The Effect of Distance in Pedestrian Detection.

The One Hundred Layers Tiramisu.

Domain Adaptation of Virtual and Real Worlds for Pedestrian Detection.

Real-time Pedestrian Detection in the GPU.