Ericssons rAppItUp: Zinkworks Journey to AI Powered rApps

At MWC 2025, Zinkworks VP of Product James McNamara joined Ericsson’s rAppItUp session to walk through Zinkworks two-year journey from early private-network experiments to a growing catalogue of AI-powered rApps, including a patented traffic prediction model now heading toward deployment with carriers in the US and EU.

Highlights

• Zinkworks’ rApp journey began with a port use case, using AI to predict device movement and traffic congestion caused by gantry cranes clustering around ships, insight that now underpins its patented Network Traffic Prediction (NTP) application, an LSTM and graph neural network model that ingests live PM/FM data to forecast cell-level congestion.
• Zinkworks has built a chained portfolio of foundational and use-case rApps, including Network Position Prediction (NPP), Quality of Service Prediction (QSP), and a Power Management Controller (PMC) that dynamically switches carriers on and off to save energy, validated through real deployment at the 2024 O-RAN Alliance PlugFest with Ericsson.
• To keep pace with rapid iteration, Zinkworks built its own no-code, GenAI-powered visual editor, using Gemini and BigQuery, that lets RAN engineers build and package new AI algorithms into deployable rApps without heavy engineering support.