Skip to content

qiyahuang72/caregiver_ai

Repository files navigation

Guardian: Privacy-First AI Caregiver

Developed for the Dell/NVIDIA Hackathon 2026

Aegis Guardian leverages edge AI to provide real-time safety monitoring for the elderly without compromising their privacy.

Key Features

  • Fall Detection: Real-time spatial analysis of human biomechanics.
  • Privacy Mode: Processes raw video at the edge and displays only anonymized "skeleton" data.
  • Hardware Accelerated: Optimized for the NVIDIA GB10 (Blackwell) architecture.

Tech Stack

  • Hardware: Dell Pro Max 10GB | NVIDIA GB10 GPU
  • AI Models: YOLOv8-Pose (Medium)
  • Library: PyTorch (CUDA 12.1), OpenCV, Ultralytics

Performance Benchmarks (GB10)

  • Inference Latency:
  • Throughput:
  • Precision:

Installation & Setup

  1. Clone the repo: git clone https://github.com/qiyahuang72/caregiver_ai.git
  2. Enter directory: cd caregiver_ai
  3. Setup Venv: python3 -m venv venv && source venv/bin/activate
  4. Install Deps: pip install -r requirements.txt
  5. Run Demo: python guardian.py

About

NVIDIA, Dell, and AI4Purpose Hackathon

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages