Mobile · 2026
Aquaponics Monitor
An IoT aquaponics system — a Flutter app paired with ESP32/ESP8266 firmware, plus on-device plant-disease detection via TensorFlow Lite.
Problem
An aquaponics setup is a small ecosystem — fish, plants, water chemistry — that drifts out of balance quietly. The grower needs eyes on it without paying for a cloud subscription or trusting a third party with their farm's data. The goal was a self-contained system: sensors on the tank, control from a phone, everything on the local network.
What it does
- Live monitoring — the Flutter app reads sensor data from ESP32 / ESP8266 nodes over the local network and shows tank conditions in real time.
- Control — actuators (pumps, etc.) are driven from the app.
- Plant-disease detection — an optional ESP32-CAM feeds an on-device TensorFlow Lite model (EfficientNet-based) that flags likely leaf disease, no server round-trip.
- Zero-config discovery — nodes are found on the LAN via mDNS, so there's no IP fiddling during setup.
Stack & build
The app is Flutter (iOS / Android / Windows). Firmware is C++/Arduino on the ESP32 and ESP8266, talking to the app over HTTP and WebSocket. Inference runs locally through TFLite — the whole system works without an internet connection, which was the point.
Why it's here
It's the most cross-disciplinary project in the set: app, embedded firmware, and an ML model shipped together, all running offline on hardware I wired up myself.