cv

Robotics Engineer

Nexuni

Building automated MIG welding systems with industrial robots

• Built automated MIG welding system with manipulator and rotary stage achieving 1mm precision using only coarse 3D models. Implemented complete perception-planning-control stack with real-time trajectory optimization that adapts to detected surface variations and ensures optimal gun angle along the weld path for high quality welds.

Internship

Siemens Lab

Developed auto-calibration systems for industrial robotics

• Developed auto-calibration system for industrial workcells that precisely determines positions of robots, cameras, and pick-and-place locations, eliminating manual setup and reducing deployment time.

Research Assistant

CMU Intelligent Control Lab

Researching multi-robot systems and human-robot collaboration

• Researched multi-robot layout optimization algorithms for optimal workcell organization and robot trajectories considering factory power constraints, workload distribution, schedule, and planning in cluttered environments.
• Explored human industrial robot control through electronic tactile skin. Applied tactile sensing and machine learning to improve human-robot collaboration and achieve safe control.
• Researched learning from physical human feedback on collaborative robots where the user expresses their intent by physically intervening the robot motion, enabling the planner to adapt to user preferences even when the environment changes.

Research Assistant

San Diego Supercomputer Center

Applied deep learning to extract causal event dependencies from text

• Designed a deep learning model that extracts causal event dependencies from text by integrating large language models with formal logic solvers to construct temporal knowledge graphs from unstructured documents.

M.S.

Carnegie Mellon University

Electrical & Computer Engineering

• Kinematics
• Dynamics & Control
• Parallel Computing
• Deep Learning
• Distributed Embedded Systems

B.S.

University of California, San Diego

Computer Science

• Advanced Datastructures
• Operating Systems
• Computer Graphics
• Computer Architecture
• Network

Optimizing Multi-Touch Textile and Tactile Skin Sensing Through Circuit Parameter Estimation

IEEE International Conference on Robotics and Automation (ICRA)

Research on improving multi-touch textile and tactile skin sensing through circuit parameter estimation.

Learning from physical human feedback: An object-centric one-shot adaptation method

IEEE International Conference on Robotics and Automation (ICRA)

Research on learning from physical human feedback with one-shot adaptation. Received Best Physical Human Interaction Paper Award.

Customizing Textile and Tactile Skins for Interactive Industrial Robots

ASME Letters in Dynamic Systems and Control

Research on customizing textile and tactile skins for interactive industrial robots.

cfs-ros

Developed a Python package implementing a kinodynamic motion planner based on the Convex Feasible Set algorithm for ROS to control humanoid robots. Available via PyPI: "pip install cfs-ros"

• Kinodynamic motion planning
• ROS integration
• Humanoid robot control