gradslam - library available!

gradslam is available as an open-source pytorch framework here

Organizing a robot learning seminar series

We are organizing a (virtual) seminar series on robot learning at Mila and the Robotics and Embodied AI Lab. Visit our seminar page for more.

Neurips workshop on differentiable vision, graphics, physics

We are organizing a (virtual) workshop at Neurips 2020 on differentiable computer vision, graphics, and physics in ML, with a star-studded speaker lineup! Visit our workshop webpage for more.

Selected to the "RSS pioneers" 2020 cohort

Honoured to be part of the RSS pioneers cohort of 2020 announced, with 27 other amazing grad students and postdocs. Check out a short video abstract of my near-term research.

Maplite awarded "Best Robotics and Automation Letters paper of 2019"

Honoured to be announced best paper of RAL, 2019, at the awards ceremony at ICRA 2020.

gradSLAM is accepted to ICRA 2020

Our paper on fully differentiable dense SLAM will be (virtually) presented at ICRA 2020

MonoLayout presented at WACV 2020.

Our paper Monolayout, tackling amodal scene layout estimation from a single image, was presented at WACV 2020. For more details, visit our GitHub repo, or watch this video abstract.

Maplite accepted to RAL; will be presented at ICRA 2020.

Our paper MapLite, tackling autonomous rural intersection navigation without a detailed prior map, is accepted for publication in Robotics and Automation Letters. It will also be presented at the International Conference on Robotics and Automation (ICRA) 2020.

Excited to launch Kaolin- a 3D deep learning library for PyTorch users

During a summer internship at NVIDIA, Toronto, I worked with Prof. Sanja Fidler’s group to build Kaolin: a 3D deep learning framework for PyTorch users.