BASc (Mechatronics/Robotics)

University of Toronto · 2017–2022

Deep Learning Road Sign Classifier for Dashcam Footage

Team of three, team lead

  • Built a complete pipeline to detect and classify road signs in dashcam videos.
  • Used YOLOv5 to locate signs and applied transfer learning on AlexNet with PyTorch to classify them.
  • Achieved 97.3% accuracy and deployed a demo interface with Gradio.

Memory Game with Microcontroller and GUI

Team of four, team lead

  • Designed and built a pattern-based memory game to enhance cognitive skills.
  • Programmed an ESP32 microcontroller to handle hardware inputs and user interactions.
  • Built a graphical user interface (GUI) to display scores and feedback in real time.
  • Implemented Bluetooth serial communication and multithreading so the GUI and microcontroller could run smoothly together.

Autonomous Robot with Mapping, Vision, and Social Interaction

Team of four, team lead

  • Programmed a robot using ROS to autonomously explore an unknown area and generate a map.
  • Added a vision module to classify images at waypoints with computer vision and machine learning.
  • Created interactive behaviors using visuals, audio, and motion so the robot could “express emotions” to users.