Robotics engineer · Edmonton, Alberta
I build robot hardware end to end, from mechanisms and circuit boards to firmware and real-time control, mostly for haptics and teleoperation.
I’m finishing an MSc in Computing Science at the University of Alberta (December 2026). My thesis is a haptic wrist for the Barrett WAM, the last stage of a project my lab has worked on since 2014. I designed it and built the final two units with Barrett Technology’s engineers, and visitors to Barrett’s booth at IROS 2026 in Pittsburgh duelled with it in a haptic lightsaber game.
The wrist, measured
Solid tags mark the main disciplines in each project; outlined tags mark supporting work.
A 3-DOF cable-driven wrist that turns a WAM arm into a 7-DOF haptic leader for a second WAM. Four prototypes at the University of Alberta, then two final units built with Barrett.
A 7-DOF WAM on a Clearpath Husky, driven from a second WAM so the operator feels what the robot touches. Built for teleoperation in disaster zones.
A one-legged hopper with a screw-cam leg that changes gear ratio along its stroke, and three reaction wheels for attitude. The controller works in simulation; the hardware build is next.
A 9-DOF quadruped built for reinforcement-learning experiments on real hardware. I rebuilt its MuJoCo model from the CAD and set up a GPU backflip task; training is in progress.
A sword-fighting game on the WAM and the haptic wrist, where blade contact pushes back through the arm. Played at Barrett’s IROS 2026 booth.
Automated bench experiments for bandwidth, friction, Z-width and virtual walls, each tested against a simulated wrist before it runs on hardware.
Our lab’s leader–follower stack for two WAMs. I worked on torque control, the gripper and haptic trigger handle, and Docker builds.
Planning CAM for 3-axis machining: a STEP part in, setups, toolpaths and LinuxCNC G-code out. Early, but it works end to end.
A Mac app that holds a smart bike trainer at a target wattage over Bluetooth, in a small overlay that floats over whatever you’re watching.
Go played on a sphere and drawn as a stereographic map, with a Monte Carlo tree search opponent. C++ and OpenGL.
A corner-balanced wheel triad, one-hop foot placement and softer landings, all in simulation.
A screw cam whose groove sets the gear ratio, and the control board built around it.
What a slow direction change revealed about the friction model.
