Saad Lahlali personal website

News: Thrilled to announce that our paper, “MVAT: Multi-View Aware Teacher for Weakly Supervised 3D Object Detection,” has been accepted to WACV 2025 in the first round! 🎉

Hello! I’m Saad, a Ph.D. candidate at Université Paris-Saclay and CEA-List, nearing the completion of my doctoral journey in 3D Perception for Autonomous Driving.

I find it incredibly satisfying to delve into the intricacies of the world around us, and I am particularly captivated by computer vision. It’s a field that allows us to replicate and understand how we as humans perceive our environment, all through the elegance of algorithms. My PhD research focuses on making 3D perception models for autonomous systems more robust, reliable, and scalable. The core challenge in this field is the reliance on vast amounts of meticulously hand-labeled 3D data, which is both time-consuming and expensive to create. To tackle this, I specialize in developing novel auto-labeling and weakly-supervised learning methods. These models are designed to learn from more accessible and cheaper data, like 2D images, drastically reducing the need for extensive human annotation while maintaining high performance.

I am always on the lookout for fresh, creative ways to tackle complex problems. Staying in the loop with the latest and greatest in the field is a priority for me, ensuring I’m always learning and pushing the boundaries of what’s possible.