Keeping fish alive is stressful, but what if Android handled the life support?

This session dives into a smart aquarium that collects sensor data, manages climate control, identifies fish via computer vision, and evaluates water toxicity using an on-device LLM, all driven by a native Android app.

Maintaining optimal aquatic conditions is high-stakes.
Tracking temperature, pH, conductivity, and turbidity lays the foundation for hardware automation. It’s a practical case study bridging raw hardware, scalable architecture, Edge AI, and delightful UI.

This session covers:
• Architecture: Interfacing sensors via MQTT and a Firebase backend.
• UI Polish: Pushing Jetpack Compose Canvas to build an interactive 2D digital twin.
• Climate Control: Securely managing high-voltage heaters and DC fans.
• ML & Edge AI: Deploying live fish identification and a compact Gemma 3 LLM via Google's LiteRT-LM for real-time telemetry analysis.

And voilà, the fish keep swimming!