Robotics
and AI Engineer
I build motion-planning, perception, control, and learning systems for robots that need to work in the real world.
Imitation Learning Based Grasping
Learning-based grasping work for robotic picking systems, connecting perception and policy outputs to real robot manipulation behavior.
Controller Framework + Adaptive Grasping
A ROS 2 control architecture for mobile manipulation, chained controllers, and adaptive grasping behavior.
Drop Detection and Drop Prevention
MuJoCo-trained drop prediction and prevention work using robot state, suction pressure, force sensing, RGB-D streams, survival analysis, CLIP, and MPC.
High-Speed Induction
High-speed robotic induction system for picking mixed parcels from clutter and placing them barcode-up onto moving sorter targets at 2000+ UPH, compared with roughly 900 UPH for human induction.
DL-Based Grasping
Deep-learning detection and grasping work for non-rigid objects, bin picking, and robotic data acquisition.
Roof and Windshield Loading
Robot-guidance software for automotive roof-panel and windshield-loading applications using industrial robots and vision sensors.
Staff Software Engineer, Robotics
Bright Machines Inc., San Francisco, CA
- Architected and led the ground-up migration of the robot platform to ROS 2, creating a scalable foundation for perception, planning, and control pipelines.
- Led development of a closed-loop rack-insertion product combining eye-in-hand computer vision for 6-DOF pose estimation with force-torque control for compliant tray insertion.
- Built end-to-end imitation learning pipelines using diffusion policy, teleoperation infrastructure, and deployable visuomotor models for contact-rich manipulation tasks.
Senior Robotics Software Engineer
Kindred AI / Ocado, Toronto, ON
- Led roadmapping and delivery for robotics software deployed on 3000+ robots, contributing to systems with impact across 1B+ picks.
- Implemented motion planning and control features including trajectory smoothing, jerk-optimized motion, tactile-informed grasping, VR controller teleoperation, and multi-robot collaboration.
- Re-architected the ROS 2 control framework and led Adaptive Grasping rollout using admittance and impedance control to reduce manipulation forces, item damage, and downtime.
- Built pipelines for object and pose detection, ML policy inference, robotic manipulation, RGBD inspection, barcode identification, CI/CD, and Hardware-in-the-Loop testing.
Software Engineer II, Robotics and Vision
EPSON Research Lab, Toronto, ON
- Adapted deep-learning detection and grasping algorithms for non-rigid objects.
- Implemented path planning and collision avoidance with ROS and MoveIt using point clouds for bin-picking applications.
- Built robotic data-acquisition and ground-truth generation systems integrating sensors, industrial robots, and motorized stages.
- Developed platforms for adaptable primitive skills with RL and IL models, earning the 2019 Epson Spotlight Award for DL-based robotic grasping work.
Robotics and Vision Software Engineer
Bluewrist Inc., Toronto, ON
- Developed robot-guidance software for roof-panel and windshield-loading applications using FANUC and KUKA robots.
- Designed and implemented vision algorithms, sensor interfacing, and sensor fusion across LMI, SICK, Zivid, and Photoneo cameras.
- Applied machine learning to a vision-guided bin-picking system where robots selected parts based on ML classification.
Emerging Tech and Innovation Engineer
General Motors, Kitchener, ON
- Led a team on vehicle-wearable integration and prototyped self-driving car diagnosis and CAN-bus communication workflows.
- Built app prototypes for stolen e-bike tracking and route selection.
- Helped ship the first GM Innovation Lab prototype to pass the D-gate.