Contributed to the perception stack for UTRA’s autonomous rover, focusing on sim-to-real vision, ROS 2 integration, and real-time deployment.

Selected Work

  • Trained and fine-tuned a YOLO-based perception model using synthetic Isaac Sim data and real rover imagery for sim-to-real autonomous perception
  • Built perception workflows for detecting lanes and terrain features from rover camera data
  • Integrated ROS 2 perception components with ZED RGB-D data to provide lane and depth information for downstream navigation

Skills

  • ROS 2
  • Python / C++
  • Computer Vision
  • YOLO
  • NVIDIA Isaac Sim
  • ZED RGB-D
  • PointCloud2