Neel Doshi

I am a Senior Applied Scientist at Amazon Robotics working on affordance generation, motion planning, software design, and control for contact-rich manipulation (e.g., Vulcan Stow).

Before this, I was a postdoctoral researcher working with Professor Alberto Rodriguez at MIT. My academic research focused on developing algorithms that leverage trajectory optimization, control/estimation theory, and mechanics to plan and execute contact-rich robotic manipulations. I completed my doctoral work with Professors Robert Wood and Scott Kuindersma at Harvard University in 2019. I developed a trajectory optimization based framework to automate the generation of closed-loop locomotion trajectories for computationally limited centimeter-scale legged robots. These tools were applied to the Harvard Ambulatory Microrobot (HAMR) to improve its performance and execute dynamic behaviors, like jumping.

Throughout my career, I have also worked on the modeling, design, and fabrication of robot end-effectors (i.e., hands and feet) that can simplify planning and control and/or be used to achieve novel capabilities (e.g., climbing). Please checkout my research and email me at nddoshi9 (at) gmail (dot) com if you have further questions.