Eric Rombokas

Timeline

1. 2001–2012

   Electrical engineering at Rice and Washington

   He earns a bachelor’s degree in electrical engineering at Rice University in 2001 and a master’s degree in electrical engineering there in 2002. In 2012, he completes a PhD in electrical engineering at the University of Washington.

2. 2012

   Learning tendon-driven hand control

   In 2012, he coauthors “Tendon-Driven Variable Impedance Control Using Reinforcement Learning” with Mark Malhotra, Evangelos Theodorou, Yoky Matsuoka and Emo Todorov. The team uses path integral reinforcement learning to learn tendon movements and feedback gains for a robotic hand performing a sliding-switch task.

3. 2013

   Postdoctoral research at Washington

   In 2013, he works as a postdoctoral researcher at the University of Washington.

4. March 2015

   Robot control with learned dynamics

   In March 2015, he, Grady Williams and Tom Daniel release the preprint “GPU Based Path Integral Control with Learned Dynamics.” Their algorithm combines machine learning of a robot’s dynamics with GPU-based sampling to plan movements in real time, and they demonstrate it on a quadrotor navigation task.

5. 2019

   Vibration feedback for prosthetic feet

   In 2019, he and Nataliya Rokhmanova coauthor “Vibrotactile Feedback Improves Foot Placement Perception on Stairs for Lower-Limb Prosthesis Users,” a study of vibration feedback for sensing foot placement on stairs.

6. October 2021

   Autoencoders for prosthetic control

   In October 2021, he coauthors and supervises “Dimensionality Reduction of Human Gait for Prosthetic Control.” The team compares principal component analysis with neural network autoencoders and finds that an autoencoder trained on entire movement trajectories improves movement classification and identification of individuals from their gait.

7. January 2023

   Gait data outside the laboratory

   In January 2023, he coauthors “A Non-Laboratory Gait Dataset of Full Body Kinematics and Egocentric Vision,” helping conceive the experiments, write the manuscript and supervise the research. The dataset contains 24.2 hours of movement recorded with wearable motion sensors, covering walking, stairs, obstacles and agility movements; one activity also includes the participants’ view and gaze.

8. As of October 2026

   Robotics and hardware at TerraClear

   In April 2026, he is named TerraClear’s Director of Robotics and Hardware after working at Cornerstone Research Group. His work focuses on making the TerraScout autonomous field scout ready for farm conditions and scaling its sensing systems. He also serves as an affiliate associate professor in the University of Washington’s departments of Mechanical Engineering and Electrical & Computer Engineering.