Sortacle

AI-powered waste bin that identifies and routes items in real time using cloud computer vision and a servo-driven sorting platform.

Project Breakdown

  • Designed the sorting tray, camera mount, motor mounts, and bearing interfaces in CAD, iterating through 6+ mechanical revisions
  • Fabricated all custom components via in-house FDM 3D printing and assembled the complete mechatronic system
  • Integrated a Raspberry Pi camera for on-device image capture and managed data flow to a remote inference service
  • Implemented a cloud-based computer vision pipeline using YOLO-World served through a FastAPI endpoint
  • Developed a servo-driven actuation system to route classified objects with repeatable and deterministic motion
  • Validated system performance through live demonstrations, earning Best Overall Project at Hack@Brown 2026

Technical Stack

  • SolidWorks (CAD & Sheet Metal)
  • Bambu Lab FDM 3D Printing
  • Raspberry Pi 4
  • PCA9685 Servo Control
  • Motors & Sensors
  • Python
  • OpenCV
  • YOLO-World (Ultralytics)
  • FastAPI (Cloud Inference)
  • Vultr (Deployment)

References