work

Some stuff I've worked on.

Research

  • COVID-Net-SSL - Improving medical diagnosis with self-supervised learning and AUC maximization
  • COVID-Net Transformer - Exploring explainability for radiography diagnosis with ViTs
  • MoE-FLAIROP - Robotic perception for grasping with Mixture-of-Experts adaptive model ensembles.
  • NRC-IRAP - Financial growth ML models for NRC’s Industrial Research Assistance Program

Robots

  • Mercedes-Benz Cars - Perception and Active Safety for L2++ capabilities (NVIDIA AV)
  • Farmonacci - ML-based perception and mapping Polymath Robotics’ trusty demo tractor
  • Tubpuller/Compound Eye - Compound Eye stereo vision and mapping ROS 2 drivers, deployed onto an agricultural vehicle (Polymath Robotics)
  • TurtleBot - Odometry, mapping, and navigation from scratch
  • AIRBOT - Grasping and manipulation from scratch with pose-based and end-to-end methods
  • Protective Stop - Open-source low-cost protective stop for robots

Hackathons

  • LeG - Turned two LeRobot arms into a pair of legs (LeRobot Hackathon SF, winner)
  • MAY - Agentic LLM code generation with self-debugging and semantic search (HackTheNorth)
  • PlayFlow - Text-to-playlist generative AI + Spotify integration + gallery (UC Berkeley AI Hackathon)

Small Projects