Benji Warburton.
EECS at Stanford — robotics, systems, machine learning
Currently working on: actuators, GNSS, robot dogs, strange vehicles.
Resume benjiw@stanford.edu GitHub LinkedIn Palo Alto, California
Experience
-
Hardware Technology Intern, Apple
Built a GNSS simulation platform for verifying signal-processing components.
-
Summer Fellow, Y Combinator
Designed robot-joint actuators with frameless motors and 3D-printed gearboxes, at roughly 6× lower cost.
-
Research Assistant, Stanford ARMLab
Working on TidyBot, an embodied-AI system that scans rooms and restores displaced objects to their original poses.
-
Teaching Assistant, CS 123
Only undergraduate on the teaching team; built labs with collaborators from Google DeepMind and mentor six robotics teams.
-
Rise Fellow, Schmidt Futures
Global Winner, 1 of 100 — designed a sub-$45 robotics education kit.
-
Data Science Intern, YASA
Analyzed electric-motor dyno tests to shorten end-of-line test cycles.
Projects
-
Electric truck restoration full blog here
Restoring a 1990s Taylor-Dunn electric truck with modern controls, drivetrain, and sensing.
-
From Text to Torque
Natural-language commands to diffusion-generated motions to PPO control policies, on a simulated Unitree G1 humanoid.
-
Compositional tool-sequence generation
Fine-tuned Qwen3 small language models for structured multi-step tool calling — over 3× better exact match on an instruction-to-program benchmark.
-
iOS RGBD streamer
An app that streams LiDAR and camera data to a computer as synchronized virtual webcams.