About Me

Hello! I’m Mansi Agarwal, an Applied Scientist at Amazon Robotics, where I build advanced AI systems that bridge the gap between cutting-edge research and scalable real-world deployment. My work spans Perception (2D and 3D), Multimodal Learning, and Embodied AI. I thrive at the intersection of research and real-world application.

At Amazon Robotics, I developed a robust multi-sensor solution for item identification, now deployed across 300+ stations and delivering measurable gains in speed, ergonomics, and cost savings. This work has been recognized as one of the team’s top contributions in 2024 and reflects my commitment to transforming ideas into products that scale.

I earned my Master’s in Robotics from the Robotics Institute at Carnegie Mellon University, where I conducted research at the Robot Perceiving and Doing Lab (RPAD) under Prof. David Held. My thesis focused on deformable object manipulation using point cloud correspondences. I hold a Bachelor’s degree from Delhi Technological University and was selected for the prestigious RISS program at CMU, supported by the SN Bose Scholarship.

My research journey has been dynamic and interdisciplinary, involving work on text analysis, computer vision, video processing, deformable object manipulation, and real-world robotics systems. This broad foundation has equipped me with the versatility to tackle complex AI challenges from multiple angles.

Outside of work, I’m a national-level swimmer, a trained Kathak dancer, and an avid reader. I bring the same dedication and discipline from these pursuits to my work in AI and robotics.