Sandro Papais

Hello! I’m Sandro Papais, a Ph.D. Candidate at the University of Toronto Robotics and Artificial Intelligence Laboratory (UTIAS), supervised by Prof. Steven Waslander. I’m also an Affiliate Researcher at the Vector Institute and a Machine Learning Perception Researcher at Zoox, where I work on advancing perception systems for autonomous driving. I co-developed the Self-Driving Cars Specialization on Coursera, a MOOC that has reached over 120,000 learners worldwide.

Before my Ph.D., I developed visual navigation and autonomy software for interplanetary spacecraft at NASA JPL’s Mobility and Robotics Group, lunar landers and eVTOLs at NGC Aerospace, and rovers at the European Space Agency’s Research and Technology Centre (ESTEC).

My work has been recognized by awards including the Qualcomm Innovation Fellowship, the Ontario Graduate Scholarship, first place at AutoDrive Challenge II, and first place at the Spaceport America Genesis Cup. My research has been published in top venues in robotics and computer vision, including a forthcoming book on motion forecasting for robotics, and was featured by U of T News for advancing tracking in self-driving cars.

I’ll be graduating in September 2026 and am currently exploring new opportunities. I’m always open to discussing research, collaborations, or ideas—feel free to reach out!

Research Interests

My research lies at the intersection of machine learning, computer vision, 3D perception, and motion forecasting, with a focus on spatiotemporal world models and end-to-end autonomous driving architectures that enable robots to perceive and act safely in dynamic environments.

Recently, I’ve been developing camera- and LiDAR-based temporal perception systems that enhance an autonomous vehicle’s ability to reason over time, maintaining object awareness through occlusions and modeling motion and object permanence. I’m the first author of ForeSight (ICCV 2025), a multi-view streaming transformer for joint 3D detection and forecasting, and SWTrack (ICRA 2024), a multiple-hypothesis tracking framework for autonomous driving. These works aim to bridge the gap between static perception and dynamic reasoning in robotic systems.