Robotics Software Engineer | Computer Vision & Autonomous Systems
I am an incoming Computer Science student in the Texas Robotics Honors Program at The University of Texas at Austin, focused on robotics software, autonomous systems, computer vision, and machine learning.
Currently, I am a Robotics Intern at NASA Johnson Space Center, where I develop perception, localization, mapping, and locomotion systems for autonomous vehicles and robots. Previously, I led software and computer vision for FRC Team 5427, building competition-ready autonomous routines, vision pipelines, and high-frequency state-estimation infrastructure.
Competition robot software developed for FRC Team 5427's 2025 season.
- Built autonomous routines, auto-alignment, and vision-assisted targeting systems
- Developed a physics-based projectile model incorporating drag, spin, and energy loss, increasing the robot's validated scoring area by 178% over a baseline parabolic model
- Implemented timestamp synchronization and latency compensation for 250 Hz odometry
- Stack: Java, WPILib, AdvantageKit, PhotonVision, CAN
A modular Java framework co-developed to support reusable robot, vision, and state-estimation systems across competition seasons.
- Helped reduce mean global translational pose error by 25%
- Designed reusable abstractions for subsystem control, localization, and computer vision
- Supported rapid iteration and debugging across multiple robot architectures
- Stack: Java, WPILib, AdvantageKit, PhotonVision
A top-down mobile simulator for the 2024 FRC game, built with realistic robot motion and collision physics.
- Simulates acceleration, inertia, collisions, game-piece interaction, intake, and shooting
- Built as a cross-platform application using Flutter's Flame game engine and Forge2D physics
- Stack: Dart, Flutter, Flame, Forge2D
A multi-task neural network that translates EEG recordings into predefined robot actions.
- Trained a joint autoencoder and classification head on 230 EEG recordings
- Classified 30 command categories and mapped predictions to robot actions
- Placed 1st among 31 intern teams at the iStart Valley capstone showcase
- Stack: Python, PyTorch, scikit-learn
- Building a ROS 2 localization pipeline that fuses visual SLAM, GPS, IMU, and wheel odometry
- Projecting YOLOPv2 road-segmentation masks into bird's-eye view and fitting road centerlines at 30 FPS on a Jetson Orin Nano
- Generating 3D occupancy maps with nvblox and Intel RealSense depth data
- Developing voxel-based scan-matching SLAM for autonomous traversal over lunar and Martian soil simulants
- Implementing inverse kinematics and gait sequencing for an 18-DOF hexapod
Languages: C++, Python, Java
Robotics: ROS 2, Isaac ROS, Gazebo, RViz, nvblox, CAN
Computer Vision & ML: OpenCV, PyTorch, scikit-learn
Tools: Git, Linux, CMake, Docker, Foxglove