SO-101 arm
I built the open-source SO-101 arm and taught it to pick and place. I teleoperate the follower with a matching leader arm to record demonstrations, then train an ACT imitation-learning policy in LeLab that runs the task on its own.
I'm Aerodynamics Team Lead for Bobcat Racing, UC Merced's Formula SAE Electric team, where I designed, optimized, and manufactured its first aerodynamic component, a carbon fiber rear wing. Off the team, I build and train robotic arms.
I lead a fifteen-person subteam, the largest in club history, across the front wing, rear wing, cooling, and body panels. I write the season plans, run our weekly work sessions, and mentor new members in CFD and composites manufacturing. I've secured over $40,000 in sponsored materials and software from partners including TCR Composites and Ansys.
I worked with the chassis team to locate the mounting hardpoints and keep the assembly inside the Formula SAE dimensional rules, then checked the mount in SolidWorks FEA across six load cases, using the aerodynamic loads from CFD.
I built an automated CFD optimization workflow on UC Merced's research computing cluster: my pipeline evaluates each wing in CFD, and DAKOTA's evolutionary algorithm picks the next size, position, and rotation of each element. It raised downforce by 50%, confirmed in a mesh-independent 3D ANSYS Fluent simulation. Across 10,000+ 3D CFD runs, I found that randomness in the optimizer changed results more than CFD fidelity, and wrote a paper on it for submission to AIAA AVIATION 2027.
I cut the cores, laid the carbon, and bagged and finished every element by hand.





This is the wing the optimization converged on. It makes 100 lbf of downforce at 60 mph.
I built the open-source SO-101 arm and taught it to pick and place. I teleoperate the follower with a matching leader arm to record demonstrations, then train an ACT imitation-learning policy in LeLab that runs the task on its own.
I designed the whole arm: the structure, the actuator assemblies, the wiring, and the software. Every joint is a NEMA 17 stepper turning a 3D-printed 19:1 planetary gearbox I designed for the torque and stiffness the arm needed, run by an ESP32 with magnetic encoder feedback. I also trained it in NVIDIA Isaac Sim to dribble a ball with reinforcement learning.

I'm from Fresno and study mechanical engineering at UC Merced, graduating in May 2028. I'm most interested in aerospace, automotive, and robotics. Outside of engineering, I'm into soccer, investing, and economics.