ART-C

Inspiration

Working in a data center comes with unique challenges. The environment is extremely loud, and temperatures can vary drastically between aisles. Efficiency is critical, as servers operate under high stress and NMC² is expected to deliver top-notch service to its customers. ART-C (Augmented Reality Technician Copilot) was created to enhance working conditions and empower data center technicians to operate more effectively on the floor.

What it does

ART-C leverages an augmented reality headset to provide technicians with a heads-up display containing critical information. It integrates with ticketing platforms like Jira to guide technicians toward open tickets closest to them. By scanning QR codes, technicians can access server metrics and logs from the past hour, reducing the need to constantly consult external screens. Given the loud data center environment, we implemented a gesture-based control system. Technicians can fetch new tasks, take screenshots, or mark tasks as complete without using voice commands or keyboards, making the workflow more seamless and hands-free.

How we built it

We built ART-C using a Meta Quest 3 AR headset for the technician display. The backend is powered by Python, which pulls Jira tickets for a given workspace and assigns tasks using a minimum cost flow algorithm. Assignment considers factors such as distance to the server, and an extra cost is added for tasks on different floors to reduce technician travel time between floors. The system dynamically updates the AR display in real-time and interacts with QR codes on servers to fetch metrics and logs, streamlining the technician’s workflow and reducing manual overhead.

Challenges we ran into

The biggest challenge was configuring the Meta Quest headset to support our gesture system. Many team members had little to no experience with Unity or the Meta Quest platform, making this a steep learning curve. Through experimentation and collaboration, we were able to implement a robust system that met our goals.

Accomplishments that we're proud of

Implemented a WebSocket connection that automatically registers and assigns tasks to technicians in real-time.
Built a dynamic AR display that updates instantly with server data and ticket information.
Developed a QR code scanning and generation system that gives technicians instant access to server metrics and logs with a simple gesture.

What we learned

We went from incredibly minimal AR/VR experience to PoC realtime solution. We learned to implement gesture detection; to create UI, model, and buttons in AR/VR; to integrate max flow to solve bi-partitate matching problems; and to navigate complex problems and multi-facted solutions.

What's next for ART-C

There’s much more we want to explore. We plan to add:

AR-enabled screenshots that can be attached directly to Jira tickets as visual documentation.
Draggable and resizable overlays for personalized AR workspaces.
Voice and gesture hybrid input, expanding accessibility and precision in interaction. Our vision is for ART-C to become an essential companion for technicians that turns complex environments into organized, intelligent spaces.