🌊 AquaFix — Autonomous Edge-AI Water Pollution Remediation
Gemma 4 Good Hackathon · Global Resilience / Climate Track
Converting smartphones into decentralized water-quality diagnostic & remediation platforms — powered by Gemma 4 E2B
Model: Gemma 4 E2B (edge, multimodal: vision + audio + text) | License: Apache 2.0 | Hardware: Zero proprietary — smartphone only
All 7 engineering tools run live below. Adjust inputs and see real-time calculations.
Phase I — Multimodal Water Quality Monitoring
Smartphone camera + Bluetooth sensor probes → pollutant identification + severity + neutralisation protocol
AquaFix v1.0 · Gemma 4 E2B · Apache 2.0 · Kaggle Notebook · GitHub
Architecture: Gemma 4 E2B (vision+audio+text) → RAG knowledge base (128K context) → 9 function-calling tools → Data Mule BT mesh → Multi-stakeholder dashboards
HuggingFace Datasets: water-potability · Tree-Species · TreeOfLife-10M · BioTrove · environmental_claims
Mobile deployment: E2B → LiteRT-LM INT4 (~1.2 GB) → React Native APK → on-device, offline, zero cloud dependency