Rui Ouyang

I am a PhD student at the Harvard School of Engineering and Applied Sciences. My PhD advisor is Prof. Roberto Rigobon, where my research will focus on machine learning applied to counter human trafficking.

I have previously worked with Prof. Robert Howe at Harvard University, Prof. Robert Platt at Northeastern University, and Prof. Edward Adelson at MIT. Prior to graduate school, I worked with Dr. Albert Wang in the MIT Biomimetics Lab (headed by Prof. Sangbae Kim). I also worked at the bioinformatics startup Curoverse and the Personal Genome Project, both under Dr. Alexander Zaranek.

In a prior life, my friend Cappie Pomeroy and I founded a startup, NarwhalEdu, focused on combining online education and hands-on hardware kits. I escaped with a BS in Mechanical Engineering from MIT. I grew up in Georgia, USA.

This website focuses on my academic work; for other work, please see my engineering portfolio. My academic pseudonym (to keep citations consistent) is Rui; in daily life, I also go by Nancy and Nao. (she/he)

Email (remove the "spam")  /  GitHub  /  LinkedIn

profile photo


My background in both mechanical engineering and computer science is reflected in my work below.

project image

Digger Finger: GelSight Tactile Sensor for Object Identification Inside Granular Media

Radhen Patel, Rui Ouyang, Branden Romero, Edward Adelson
International Symposium on Experimental Robotics, 2020
arxiv / website /

Work on a novel Gelsight design customized to a specific application. Gelsights are visual-tactile sensors used to give robots a sense of touch. The Digger Finger is a miniaturezed, custom-shaped sensor specialized for digging in sand and other granular media, which has rarely been explored with robotic touch before. I helped prototype the physical design as well as the initial shape detection and annotation algorithm which used a YOLO-OBB model.

project image

Low-Cost Fiducial-Based 6-Axis Force-Torque Sensor

Rui Ouyang, Robert Howe
IEEE International Conference on Robotics and Automation, 2020
arxiv / slides / website /

A novel six-axis force/torque sensor. The sensor is low-cost (<$50) compared to other sensors which can cost up to tens of thousands of dollars (e.g. some ATI sensors). The sensing component itself uses simple webcam and a printed paper tag, allowing for ease-of-use (just plug in the USB). The sensor is also quite small. Design files and code are open-source (see website).


project image

I received the Google Women Techmakers Scholarship in 2017.

Other Projects

Classwork and unpublished work below. Please see my engineering portfolio for more details / more projects. I also keep an informal blog at orangenarwhals.

project image

Throwdini: Throwing with Orientation by Reinforcement Learning of Physics Model Errors

Rui Ouyang, Ondrej Biza, TBA
video /

(Unpublished work) I implemented robotic arm system capable of throwing with orientation (see realtime video for example). A simple ballistics model is used to calculate initial throw parameter. The residual term is calculated with a learned model on real-life data, and accounts different grasp locations along the axis of the object.

project image

6.832: Inertia Wheel Inverted Pendulum

Rui Ouyang, Ashwin Krishna
MIT Underactuated Robotics
video / slides / report /

Using a flywheel to invert a pendulum. In simulation, we derive the equations of motion and apply LQR and region of attraction analyses for our system. We also build a hardware system from scratch and implement swingup and stabilization using PD and bang-bang control. I did the analysis and controls and co-built the hardware system.

project image

CS279r: PaperSignals Parameterized

Rui Ouyang, Laura Zharmukhametova, Jamelle Watson-Daniels
Harvard PL/HCI Graduate Seminar
code / report /

PaperSignals are build-it-yourself printable robots that a user can control with their voice. This study presents an open-source browser-based user interface (written in Python) aimed at allowing users to easily customize existing Papersignal templates. I cowrote the Flask webapp and built the physical prototypes.

project image

CS223: Efficient Algorithms for Detecting & Classifying Credit Communities

Rui Ouyang, Juspreet Singh Sandhu, Mark York
Harvard Random Processes and Algorithms
code / report /

How can creditworthiness of individuals be established in the absence of reliable data? We investigated the propagation of trust scores when individuals are asked to recommend other borrowers, and the resulting error and discovery rates. I wrote the simulation code for our results using NetworkX.


I am also an activist and a feminist.

project image

Opinion: Wearing masks must become mandatory

Rui Ouyang

website / pdf /

Wrote op-ed about mandating masks in my home state, Georgia. Written in mid-May, the article was published in the Atlanta Journal Constitution in July when rising cases caused dozens of states to mandate masks. Two weeks later, our State Governor Kemp sued Atlanta City Mayor Bottoms to end her city-wide mask mandate.

project image

Opinion: Shut Ga. down now to save lives, economy

Akshat Pujara, Mike Callahan, Rui Ouyang

website / pdf /

Co-wrote op-ed supporting a shutdown of my home state, Georgia. I compared our state’s situation to other countries such as Italy, and we explained how the economic and healthcare interests were aligned in the long run. The article was published online in the Atlanta Journal Constitution shortly before the state shutdown on April 1st.

project image

#ShutdownGA Campaign

Rui Ouyang

website / pdf /

I created a petition to shut down my home state (Georgia) and got 80k+ signatures in 2 weeks. I sent it to my governor and federal representatives. I also did GA-specific hospital capacity analysis and emailed it to over >200 state legislators and city mayors. I made a website to coordinate volunteers.

Design and source code from Jon Barron's webdsite. as modified by Leonid Keselmam.