Projects

A selection of research and engineering work spanning robotics, machine learning, computer vision, and software development.

project

DYNAMO — Multi-Robot Transport

My M.S. thesis research (awarded distinction). Developed dynamic admittance parameterization for non-prehensile cooperative transport using mobile robots with fixed manipulator arms. Combines hybrid admittance-impedance control with optimal coordinated planning for robust payload manipulation without grasping.

ROS C++ Python SLAM Admittance Control Motion Planning
project

RL Quadrotor Control

End-to-end multi-behavior policy for quadrotor control using modern deep reinforcement learning techniques. Trained in NVIDIA Isaac simulation environments with sim-to-real transfer capabilities, enabling complex autonomous flight behaviors from a single learned policy.

PyTorch Isaac Sim Reinforcement Learning CUDA Python
project

Real-Time Object Tracking Pipeline

Computer vision pipeline for simultaneously tracking small objects across multiple video feeds. Developed at Sandia National Laboratories, combining custom detection models with multi-camera synchronization for reliable real-time object localization.

OpenCV Python Computer Vision Real-Time Systems
project

Physics-Informed Constraint Estimation

Trained a physics-informed deep neural network and novel localization algorithm to rapidly estimate mechanical constraints of objects. Enables robust autonomous robotic manipulation by understanding physical properties before contact.

PyTorch Physics-Informed ML Robotics Python
project

RGBD Detection & Localization

Novel real-time RGBD-based detection, segmentation, and localization ROS package. Integrated custom YOLO models with OpenCV filters and CUDA acceleration, achieving accurate centroid estimation with sub-frame inference on a single GPU.

ROS YOLO OpenCV CUDA C++ Python
project

Gesture Recognition Testbed

Designed and built an AprilTag and PID-based outdoor experimental testbed for benchmarking gesture recognition algorithms on mobile robots. Evaluated reliability and accuracy of hand-sign recognition across multiple robot platforms in outdoor environments.

ROS AprilTag PID Control Python Mobile Robots
project

Productivity Web App

Full-stack web and mobile application for habit tracking. Built with Firebase backend and Flutter frontend, featuring smart habit adherence mechanisms. Included project planning, task management, and milestone tracking for collaborators.

Firebase Flutter Dart Full-Stack
project

Legislature Summarizer

AI-powered web application that consolidates and summarizes government legislature, making policy accessible and searchable. Built during the 2025 Lobo Hackathon, where it placed fourth.

AI Web App NLP Hackathon