User builds real-time poker game with Reachy Mini and Eliza using Gemma 31B

A developer created a live poker experience featuring robots Reachy Mini and Eliza as part of the Cerebras x Gemma 24hr Hackathon. The system uses an orchestrator to guide the game and resolve showdowns without manual buttons.

• Vocalized actions allow players to interact like in a real game.
• A webcam tracks table cards while robots use their own cameras for private hands.
• Gemma 31B serves as a vision-language model for real-time card tracking.
• Inference is provided online via Cerebras.

The project demonstrates the capability of using large models for real-time visual tasks in interactive robotics.