Liam Paull
Liam Paull is a roboticist at Université de Montréal and Mila whose work covers robot learning, mapping and autonomous driving. After a PhD on underwater vehicles at the University of New Brunswick in 2013, he co-led MIT's Duckietown course in 2016 and coauthored the paper on the AI Driving Olympics, first held at NeurIPS in December 2018. In 2019 he became a Canada CIFAR AI Chair and coauthored gradSLAM, a differentiable approach to simultaneous localization and mapping, and he received an IEEE Robotics and Automation Letters Best Paper Award in 2020 and the CS-Can/Info-Can Outstanding Early Career Computer Science Researcher Award in 2023. He is an associate professor at Université de Montréal, a core academic member of Mila and director of the Duckietown Foundation, and he co-leads the Montréal Robotics and Embodied AI Lab (REAL).
- Now at
- Mila, Duckietown Foundation and Montréal Robotics and Embodied AI Lab
Timeline
-
2013
PhD on underwater vehicles
He completes a PhD at the University of New Brunswick in 2013, co-supervised by Mae Seto, Liuchen Chang and Howard Li. The doctoral work focuses on robust and adaptive planning for underwater vehicles.
-
2016
Duckietown at MIT
He works on simultaneous localization and mapping (SLAM) for underwater robots as a postdoc in John J. Leonard's Marine Robotics Group at MIT. In 2016, as a CSAIL postdoc, he co-leads MIT's new Duckietown course with research scientist Andrea Censi, which aims to build a fleet of 50 duckie-adorned robo-taxis that navigate with one on-board camera and no pre-programmed maps. He also uses the platform to prototype algorithms for CSAIL's $25 million collaboration with Toyota on autonomous cars.
-
2017
Moving from MIT to Montréal
Having led the CSAIL/Toyota Autonomous Car Project as a research scientist in Daniela Rus's Distributed Robotics Lab at MIT, he teaches at Université de Montréal by the fall of 2017, when classes at Montréal, ETH Zürich, TTI Chicago and NCTU run as one coordinated Duckietown course. He is first author of the 2017 ICRA paper that presents Duckietown as an open, inexpensive and flexible platform for autonomy education and research.
-
December 2018
The AI Driving Olympics
In December 2018, the first AI Driving Olympics (AI-DO), a Duckietown-based competition for machine learning and AI in mobile robotics, takes place at the NeurIPS conference in Montreal. He is one of the 17 authors of the paper that introduces it.
-
2019
A Canada CIFAR AI Chair
CIFAR appoints him a Canada CIFAR AI Chair at Mila in 2019 and renews the chair in 2025. His group builds representations of the world for simultaneous localization and mapping, models uncertainty, and develops workflows to teach robotic agents new tasks through simulation or demonstration.
-
2019–2020
Differentiable SLAM with gradSLAM
He coauthors gradSLAM with Krishna Murthy Jatavallabhula, Soroush Saryazdi and Ganesh Iyer, first posted to arXiv in October 2019. The ICRA 2020 paper poses dense SLAM as differentiable computational graphs, so that gradients flow from 3D maps to 2D pixels.
-
2020–2023
Two research awards
He receives a Best Paper Award from IEEE Robotics and Automation Letters in 2020 and the CS-Can/Info-Can Outstanding Early Career Computer Science Researcher Award in 2023.
-
As of October 2026
Professor at Université de Montréal
He is an associate professor at Université de Montréal, a core academic member of Mila and co-leader of the Montréal Robotics and Embodied AI Lab (REAL), and he directs the Duckietown Foundation. In May 2026, Mila announces a partnership with Torc Robotics, and he welcomes new collaborations between academic labs and autonomous-vehicle companies.
Career
- Since 2019MilaCanada CIFAR AI Chair
- NowUniversité de MontréalAssociate Professor
- NowDuckietown FoundationDirector
- NowMontréal Robotics and Embodied AI LabCo-leader
- Until 2017Massachusetts Institute of TechnologyPostdoc
- MIT CSAILPostdoc
Education
- 2013University of New BrunswickPhD
Sources and credits
Photo: We have no published photo of Liam Paull yet. Send us a free-licensed photo.
- Liam Paull | Mila
- Home | Liam Paull
- Self-driving cars, meet rubber duckies | MIT News
- A brief history of Duckietown - Duckietown
- Liam Paull – CIFAR
- The AI Driving Olympics at NeurIPS 2018
- gradSLAM: Automagically differentiable SLAM
- Torc Robotics Partners with Mila in Institute's First-Ever Autonomous-Trucking Collaboration to Advance Physical AI | Mila



