PersonalPublic2024

Waeky-Waeky

Built using Tauri and Flask to combat early-onset narcolepsy, Waeky-Waeky monitors alertness during intense study sessions. Struggling to stay awake while reviewing course materials, I engineered this background utility to detect drowsiness and trigger immediate wake up alerts. This project was pivotal in saving my grades, enabling me to pass my second year of college when I otherwise would have failed the semester. It uses optimized facial landmark tracking and lightweight computer vision pipelines to run continuously in the background.

Waeky-Waeky

About This Project

Combating severe, early-onset narcolepsy that caused me to fall asleep during study sessions, directly risking failure in my second year university semester.

Developed a background application using Tauri, Rust, and Flask. Implemented real time facial landmark tracking to calculate eye-aspect ratios via OpenCV, triggering wake up events.

Directly saved my academic standing; the utility allowed me to pass my second year college courses by successfully keeping me awake during study blocks.

Role

Machine Learning Engineer

Year

2024

Status

Public

Type

Personal

Technology Stack

PythonOpenCVFlaskTauriReactRustTailwind CSS

Project Story

The Challenge

Combating severe, early-onset narcolepsy that caused me to fall asleep during study sessions, directly risking failure in my second year university semester.

The Approach

Developed a background application using Tauri, Rust, and Flask. Implemented real time facial landmark tracking to calculate eye-aspect ratios via OpenCV, triggering wake up events.

The Outcome

Directly saved my academic standing; the utility allowed me to pass my second year college courses by successfully keeping me awake during study blocks.

Insights & Takeaways

Highlights

  • Directly resolved a severe personal health hurdle (narcolepsy), saving my second year grades.
  • Uses optimized facial landmark tracking algorithms to detect sleep onset.
  • Runs continuously as a lightweight background daemon combining Tauri and a Flask process.

Challenges

  • Achieving 100% sleep detection accuracy without lag or CPU spike on study hardware.
  • Calibrating eye-aspect thresholds to differentiate between downcast reading gazes and sleep.

Lessons Learned

  • Witnessed the tangible, life changing impact that personalized software can have on individual health and academic success.

Related Work