Raven

Too Hot
View of the Sensors
Regular Status
Step-by-step instructions

Inspiration

Data center technicians work in environments that are loud, hot, and often unpredictable. Their safety and efficiency depend on how quickly they can adapt to changing conditions while managing complex tasks and following dense technical documentation. Our personal device helps them stay safe, focused, and efficient, combining real-time hardware sensing with AI-powered workflow support.

What it does

Our project is a wearable assistant for data center technicians powered by the M5GO IoT device and the Gemini API.

It helps technicians:

Stay Safe-

-Tracks temperature, humidity, and sound level (decibels) in real time.

-Displays a red warning light when heat or noise exceed safe thresholds.

-Monitors prolonged high-decibel exposure and warns accordingly.

Stay Efficient-

-Uses Gemini API to summarize complex server documentation (like the 96-page NVIDIA GB200 NVL72 guide) into clear, step-by-step instructions.

-Each step can be navigated with simple forward/backward buttons on the device, reducing cognitive load and letting technicians stay focused on the task.

-When a ticket is complete, the technician can mark it done directly from the interface.

How we built it

Hardware: M5GO IoT device (ESP32-based) with environmental sensors and LEDs.

APIs:

-Gemini API for summarizing long technical documentation and generating clear procedural guides. -Integrated via a Python/Flask backend.

Software:

-Custom firmware written in Python for real-time monitoring.

-Dashboard with streamlit for logging sensor data and tracking ticket progress.

Data:

Used live sensor data from the M5GO. Summarized the official NVIDIA NVL72 GB200 96-page document to test Gemini’s ability to convert dense technical content into safe, actionable workflows.