About me
I build intelligent, deployable robotic systems. Currently, I'm an Applied Scientist at Amazon Robotics, where I develop and scale learning-based solutions for contact-rich manipulation. My work leverages reinforcement learning and imitation learning to solve complex bimanual manipulation challenges at Amazon's scale, bridging the gap between cutting-edge research and high-impact industrial applications.
My path here was a bit all over the map, in a good way. During my PhD at the University of Illinois, I spent a lot of time on sequential decision-making problems and applied reinforcement learning to everything I could find—from optimizing crop management systems to teaching rovers how to navigate extra-terrestrial terrains. That rover work naturally led me to sampling and manipulation tasks, which is where I really got into contact-rich physics problems—figuring out how to get robots to reliably grasp, move, and manipulate objects in uncertain environments. It's this combination of large-scale learning and contact-rich manipulation background that I bring to my work today.
As much as I hate to admit it, I struggled to keep this page up to date over the years. At some point, I decided to just nuke everything and keep it simple.
I genuinely love chatting with people about literally anything—whether you need advice, want to collaborate, have some wisdom to share, or just feel like saying hi. Drop me a line at [email protected] or find me on socials. I promise I'm way better at responding to emails than updating websites.
My path here was a bit all over the map, in a good way. During my PhD at the University of Illinois, I spent a lot of time on sequential decision-making problems and applied reinforcement learning to everything I could find—from optimizing crop management systems to teaching rovers how to navigate extra-terrestrial terrains. That rover work naturally led me to sampling and manipulation tasks, which is where I really got into contact-rich physics problems—figuring out how to get robots to reliably grasp, move, and manipulate objects in uncertain environments. It's this combination of large-scale learning and contact-rich manipulation background that I bring to my work today.
As much as I hate to admit it, I struggled to keep this page up to date over the years. At some point, I decided to just nuke everything and keep it simple.
I genuinely love chatting with people about literally anything—whether you need advice, want to collaborate, have some wisdom to share, or just feel like saying hi. Drop me a line at [email protected] or find me on socials. I promise I'm way better at responding to emails than updating websites.
© 2025 Pranay Thangeda. All rights reserved.