From February 2024 through May 2025, I worked as a research assistant (promoted upon graduating to research engineer) at UT Austin's Nuclear and Applied Robotics Lab under Dr. Mitch Pryor. While at NRG, I was the lead engineer overseeing our lab's contribution to the robotic perception stack for the USSF's Space Strategic Technical Institute for In-Space Operations.

​

Our goal was to determine how to capture a 3D reconstruction of a satellite while that satellite was taking evasive actions to avoid being scanned. I was the lead research engineer and project manager in developing a system that uses a next-best-view planner (https://arxiv.org/abs/2203.10113) to find the camera poses a robot needs to reach in order to maximize new information gained about the target — i.e., the system minimizes the number of camera poses required by maximizing the amount of information gained with each new viewpoint. Once we had generated the reconstruction of our target, we ran a volumetric grasping network (https://arxiv.org/abs/2101.01132) on the reconstruction to locate grasp points on the target satellite. As part of this project, I also upgraded the software stack for SCRAT, our custom-built holonomic robot that served as the chase vehicle for this demonstration. I upgraded it from ROS 1 to ROS 2.

​

As lead engineer for this project, I managed a group of four graduate students and two undergraduates while also being responsible for my own contributions. I set timelines for deliverables, led meetings, and ensured that everyone's individual components integrated successfully for our final demo, which I presented to a consortium of universities and customers in April 2025.

​

Below I've attached a video of SCRAT after it had been upgraded to ROS2, a videos of our final integrated demo and generated reconstruction, and a picture our lab took with a Robonaut R2 at the year 1 progress update meeting at TAMU.