Natey Kim

Human–Robot Interaction Research Engineer @ UPenn GRASP Lab

Photo of Natey (1Korok ProfPic.png)

I am a research engineer at Penn’s GRASP Laboratory working at the intersection of assistive robotics, human movement, and human–robot interaction. I earned an M.S.E. in Robotics and a B.S.E. in Bioengineering from the University of Pennsylvania.

My research spans upper-limb exoskeletons, EMG-informed musculoskeletal digital twins, human-in-the-loop control, machine learning for injury assessment, and socially assistive robots. I build systems that translate neuromuscular intent into adaptive, intuitive, and clinically meaningful technologies.

Alongside my work at Penn, I am a research resident at Maingen, where I study machine-learning methods for robotic end-effector design, and a founding mechanical engineer at Tadashi Robotics, where I am helping translate the Ember social-robotics platform into a rehabilitative product.

Background

Research Experience

GRASP Laboratory, University of Pennsylvania

Graduate Research Assistant → Research Engineer

Aug 2023 – Present
Philadelphia, PA

Develop musculoskeletal digital twins, real-time control and signal-processing systems, and machine-learning methods for upper-limb exoskeletons. I also develop high-density surface EMG hardware and analysis pipelines for movement assessment, rehabilitation-exercise classification, and Achilles tendinopathy detection.

Upper-limb exoskeleton project HD-sEMG and Achilles classification project

Research Output

Publications

Maggie Wagner*, Sanghyub Lee*, Nathaniel Kim*, Michael North*, Nadia Figueroa, Josh Baxter, and Flavia Vitale

Latent HD-sEMG Representation Learning for Achilles Tendinopathy Assessment

Pending submission to IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2026.

Sanghyub Lee, Nathaniel Kim, Dhyey Shah, Jason Chen, and Nadia Figueroa

Musculoskeletal Digital Twin Modeling for Biomechanical State Estimation During Weighted Elbow Flexion

Under review for the 2026 IEEE-RAS International Conference on Humanoid Robots (Humanoids).

* Equal contribution.

Portfolio

Research Projects

Hybrid Simulation and EMG-Controlled Upper-Limb Exoskeleton
Penn GRASP Laboratory · Philadelphia, PA • Jun 2025 – Present
Developed a MuJoCo musculoskeletal digital twin using inverse dynamics, PD control, and quadratic programming, reproducing measured elbow kinematics with a mean tracking RMSE of 1.866°. Built a causal temporal convolutional network that predicts subject-specific biceps and triceps EMG for unseen controller configurations at real-time-capable latency. Integrated surface EMG, OptiTrack motion capture, motor torque, and exoskeleton encoder data across participants, external loads, and controller configurations.
  • Digital-twin and musculoskeletal simulation comparison
EMG-Based Achilles Tendinopathy Classification
Penn GRASP Laboratory · Philadelphia, PA • Aug 2023 – Present
Developed a real-time high-density EMG collection pipeline using custom 3D-printed hardware and laser-cut MMX/silicon fabric electrodes. Built a 3D variational autoencoder to quantify injury severity, assess laterality, and classify rehabilitation exercises. Detected Achilles tendinopathy in an average of 20 of 21 held-out evaluations and classified healthy, unilateral, and bilateral pathology in 18 of 21.
Images withheld pending publication.

Background

Work Experience

Maingen (Y Combinator)

Research Resident

Jul 2026 – Present
Remote

Research machine-learning methods for generating robotic end-effector CAD geometries and build simulation and evaluation pipelines that test geometric validity, mechanical performance, and task-specific engineering constraints. My benchtop project develops an autonomous pipeline that compiles generated parametric designs into physically grounded MuJoCo simulations and iteratively repairs them using deterministic mechanical, collision, and robustness checks.

Tadashi Robotics

Founding Mechanical Engineer

Aug 2025 – Present
Philadelphia, PA

Serve as the sole engineer developing Ember as the social-robotics research platform moves into an early-stage rehabilitation startup. I lead the evolving mechanical design, embedded architecture, sensing and communication hardware, system integration, prototyping, and LLM-based patient interactions.

Ember social robotics project

Portfolio

Work Projects

Ember Social Robotics Platform
Tadashi Robotics · Philadelphia, PA • Aug 2025 – Present
Translating Ember from an academic social-robotics research platform into an early-stage product for rehabilitative patient support and naturalistic human–robot interaction. Designed the mechanical enclosure and embedded architecture across wireless communication, audio I/O, vision, display, power, and motor-control hardware. Developed LLM-based interaction logic for reminiscence dialogue and emergency guidance and supported human-subject research submissions.
  • 1st prototype of Ember

Portfolio

Other Projects

Verdigris: Assistive Ultrasound Guidance for Lumbar Puncture
University of Pennsylvania · Philadelphia, PA • Aug 2025 – May 2026
Co-developed an assistive medical device using ultrasound-based bone detection and real-time angle feedback to guide lumbar-puncture needle trajectories. Designed the enclosure and interface, integrated the ultrasound transducer, motor, angle sensor, and control electronics, and tested acoustic acquisition on tissue and spinal phantoms. Validated angle sensing and LED feedback within 1° across controlled targets; the prototype achieved 80% optimal needle-path detection and 95% bone-detection accuracy.
  • Rotating CAD model of the Verdigris enclosure
Rehabilitative Driving Simulator
Kingston, Jamaica • Jan 2023 – May 2023
Collaborated with engineers and clinicians in Jamaica to develop a Unity-based driving simulator for pediatric patients with hemiplegia. Created custom steering and pedal controls and fabricated the supporting mechanical, electrical, and sensor-integrated components.
  • Patient using the device for the first time!
Social DinoBot
Philadelphia, PA • Aug 2024 – Dec 2024
Designed a mobile dinosaur social robot to aid gait rehabilitation for an ischemic stroke patient. Fabricated wooden chassis for lightweight frame and modular assembly. Implemented various modes of communication for assistance with aphasia and speech rehabilitation.
  • Social DinoBot rehabilitation prototype

CV

Open CV

If the PDF viewer does not load, download the CV.

Curious to see who I am beyond the work?

Explore Beyond the Work