Selected work

Robots built from mechanics to control.

Six research projects spanning high-payload climbing, dexterous and dynamic manipulation, real-time perception, model predictive control, and multi-robot systems.

Six selected projects

Research

Robot hardware, control, perception, and learning.

ICRA 2026 2023–present

High-payload wall climbing

Mechanical design · Control

Passive suction enables an 8 kg payload on plaster and glass.

My role Mechanism design & payload testing

Paper Website

Project details

FREE Lab · University of Michigan

Problem

Carry payloads on vertical surfaces using passive bistable suction cups.

My contribution

I designed the passive bistable-suction wall-climbing robot and evaluated its payload capability on plaster and glass.

Result

Demonstrated an 8 kg payload on plaster and glass; first-author ICRA 2026 paper and a 2025 U.S. provisional patent application.

Mechanism design
Suction
Hardware validation

ICRA 2027 submission May–Aug. 2026

Steerable & reactive grasping

Manipulation · Learning

Modular three-finger grasping with reactive reaching and tactile feedback.

My role Grasping framework & hardware validation

Paper Website

Project details

Biomimetic Robotics Lab · MIT

Problem

Connect object geometry, reactive reaching, and tactile feedback in a modular grasping framework.

My contribution

I developed the three-finger grasping framework, combining geometry-based affordances, reactive reaching, and tactile/proprioceptive reinforcement learning, and demonstrated it on hardware.

Result

Hardware demonstrations on seen and unseen objects; released as a first-author arXiv preprint and submitted to ICRA 2027.

Reactive manipulation
Reinforcement learning
Tactile sensing

Ongoing research May–Aug. 2026

MIT Table Tennis Robot V2

Perception · Control · Learning

A custom 220 Hz perception stack for MIT's second-generation table-tennis robot.

My role Real-time perception & system integration

Project details

Biomimetic Robotics Lab · MIT

Problem

Give a high-speed table-tennis robot reliable ball state estimates and connect them to learning- and model-based controllers.

My contribution

I implemented 220 Hz YOLO-based ball localization and integrated it with the hardware system, assisting with MPC controller deployment and testing for the robot's table-tennis skills.

Result

Integrated real-time ball localization with the robot's control stack for hardware testing across dynamic table-tennis skills.

Real-time perception
Model predictive control
System integration

Science Robotics 2026 May–Aug. 2024 & 2025

AthenaZero

Dynamic manipulation · Robot hardware

Low-inertia bimanual hardware for throwing, catching, and batting at human-comparable speeds.

My role Hardware intern · actuator testing & assembly (not a paper author)

Paper Website

Project details

RAI Institute · formerly Boston Dynamics AI Institute

Problem

Build a bimanual manipulator with low effective inertia and high control authority for fast, contact-rich tasks at human-comparable timescales.

My contribution

Across two summer internships, I built an automated Python motor-dynamometer test suite, characterized actuator friction and torque–velocity behavior, designed a 1-DOF camera mount, and assembled 40 quasi-direct-drive actuators using custom test fixtures.

Result

The resulting platform achieved a human-comparable effective endpoint mass and demonstrated throwing above 30 m/s, with catching and batting above 14 m/s.

Actuator characterization
Hardware integration
Test automation

Research project 2026

FLINT feedback control

Model predictive control · Simulation

MPC-based feedback control for randomized robot models.

My role MPC feedback control & simulation

Project details

University of Michigan

Problem

Track optimized motions when robot geometry and dynamics vary between systems.

My contribution

I developed and evaluated the MPC feedback-controller workflow in FLINT, using randomized robot models to test closed-loop trajectory tracking across changing mechanisms.

Result

Demonstrated feedback-control rollouts across randomized two-link robot models in simulation.

Model predictive control
Feedback control
Robot simulation

Research project

Coordinating fixed-wing UAV formations

Simulation · Control

Simulation and control for interconnected fixed-wing formations.

My role Simulation & formation controllers

Project details

FREE Lab · University of Michigan

Problem

Simulate and coordinate interconnected fixed-wing aircraft in formation.

My contribution

I developed the simulation and control stack for interconnected fixed-wing UAV formations.

Result

Implemented Circle/Flying-V and waypoint-following controllers in the simulation and control stack.

Simulation
Formation control
Waypoint following

Additional work

Engineering projects

Hands-on systems and competition robotics.

Design project Jan.–Apr. 2025

P-Wheel: fluid-powered flow sensing

Design · Embedded systems

Fluid-driven flow sensing with real-time LED and LCD feedback.

My role Mechanism, sensing & embedded feedback

Project details

University of Michigan

Problem

Measure stream velocity without pumps or wall power and provide interactive feedback.

My contribution

I designed the fluid-driven paddle wheel and encoder sensing system, and built an LED and LCD feedback loop that gamifies aim against cohort high scores.

Result

Our team ranked first in the cohort and was one of four selected to pitch to venture capitalists.

Mechanical design
Sensing
Embedded feedback

Competition robotics 2012–2023

Competition robotics

Robot design · Software · Leadership

Competition robots built through years of design, software, and testing.

My role Team captain · Robot design · Software

Video

Project details

The Brainstormers

Problem

Design, program, and refine competition robots while leading a team.

My contribution

As team captain, head robot designer, and software lead for The Brainstormers, I led robot design and software and mentored more than 100 teams.

Result

Team highlights include a 2022 World Championship finalist finish, six state championships, and a 2023 World Championship Innovate Award second-place finish.