LORENZ ATTRACTOR · σ=10 ρ=28 β=2.667 (?)

// why I'm into chaos

I care about Lorenz and chaos theory because they show how small changes at the micro level can create very different behavior at the macro level. That idea fascinates me because it means the whole system can become unpredictable even when the individual rules are simple.

HI, I'M

Bijan Mehr

▌

What I'm doing now → Get in touch

σ 10.0 ρ 28.0 β 2.67

About.

I'm Bijan Mehralizadeh, a first-year Computer Science PhD student at George Washington University with a background in mechatronics engineering. I've tried my hand at being a "full-stack engineer" (a term a lab mate coined for someone who wants to build everything), but right now I'm laying the groundwork for research on making autonomous systems — drones, robots, IoT devices — more resilient and explainable. It's early days, but I'm focused on how these machines can handle failures or attacks without falling apart. When I'm not in class or the lab, you'll find me taking gadgets apart, gaming, or hiking as an outdoor nerd — blissfully forgetting all the bugs I left behind.

Education

Ph.D. Computer Science — Jan 2025 – Present
M.S. Mechatronics Engineering — 2017 – 2021
B.S. Mechatronics Engineering — 2012 – 2017

Skills

PythonC/C++MATLABSimulink ROS/ROS2GazeboMoveItRViz TensorFlowPyTorchOpenCVMediaPipe Scikit-learnKeras

Now.

What I'm building and digging into these days.

Live

Teaching

PID control, made interactive

A teaching demo I built for an autonomous-systems course — drag the PID gains and watch a car steer through a lane change, in real time.

Try the demo →

In progress

Research

Multi-agent systems

Cooperating agents that hold up under failure and attack.

Planned

Learning

ML & RL for control

Picking up reinforcement learning and applying it to control.

Ongoing

PhD

@ GWU · Systems Security Research Group

First year — resilience & security of autonomous systems.

Journey.

STAGE 01/14YEAR 2012

▮

← / → TO MOVE

Publications.

The honest one-liners. Mostly autism-screening robots, back in my mechatronics days.

ICROM · 2023

Can a pile of sensors catch autism early? Our first real swing at it — I helped build the multi-modal screening rig, and the early signal looked promising.

Multi-modal ASD screening system, a preliminary study the formal version →

Sustainability · 2023

We stuffed a toy car full of sensors and watched kids play with it — turns out how they drive it carries real autism-screening signal. I built the sensorized car + the feature pipeline.

A Sensorized Toy Car for Autism Screening Using Multi-Modal Features the formal version →

Frontiers in Robotics & AI · 2023

A whole room of social robots that help kids practice reading and showing emotions. I helped build the system that choreographs the robots.

Fully robotic social environment for teaching and practicing affective interaction the formal version →

ICSR · 2020

The "why bother" paper: making the case that robots are a surprisingly good, low-pressure space for autism therapy. The argument that kicked off all the rest.

Robotic Social Environments: A Promising Platform for Autism Therapy the formal version →

Full record on Google Scholar →

Let's talk.

LAT / LON38.9072°N · 77.0369°W

STATIONGWU · Washington, DC

EMAILbijan.mehralizadeh@gwu.edu

LINKEDIN/in/bijanmehr →

SCHOLARGoogle Scholar →