Halo – A Virtual Caregiver for Alzheimer’s Patients

Inspiration

The inspiration behind Halo comes from my own family. My grandparents struggle with memory deterioration, and my mom and I are often away due to school and work. This has caused frequent worry about their emotional and physical well-being. Recognizing that this is not just a local problem but a global one, I built Halo to help families support their elderly loved ones emotionally, even from a distance. In the future, I hope to integrate Halo directly into mobile operating systems to make it even more accessible.

What I Learned

  • Built a full-stack application from scratch using Vite.
  • Learned to integrate external APIs for TTS (Text-to-Speech) and STT (Speech-to-Text).
  • Improved my system design skills and front-end component architecture.
  • Gained experience working with asynchronous programming and voice interaction logic.

How I Built the Project

  • Started by scaffolding the front end with Vite.
  • Designed the UI to be accessible for elderly users using large fonts, clear navigation, and soft color palette.
  • Created React components for:
  • Medication reminders
  • Memory prompt delivery
  • A caregiver interface
  • Built the backend with Express.js and MongoDB to store user data, memory prompts, and medication schedules.
  • Integrated TTS and STT APIs for voice input/output.
  • Added local and scheduled reminders using a custom reminder engine.

Challenges I Faced

  • Text-to-Speech and Speech-to-Text: It was challenging to ensure that speech recognition was accurate across different environments and accents.
  • Implemented fallback checks to prevent reminders from being skipped due to voice misinterpretation. The biggest technical issue was ensuring medication reminders were 100% reliable and I had to create a fallback system in case time-based triggers failed due to app inactivity.
  • Designing AI interactions that were simple, empathetic, and not overwhelming for elderly users was difficult. I tried to avoid overly complex NLP models in favor of clear, friendly prompts.

Halo is more than a project, it is a deeply personal solution aimed at preserving dignity, memory, and connection for Alzheimer’s patients. By combining voice interaction, memory reinforcement, and structured care, Halo helps families stay close, even when apart. In the future, I hope to integrate hardware features like smartwatch syncing and camera-based behavior tracking for safety alerts.

Built With

Share this project:

Updates