ECK
Back to Projects
ai

FisheryAI

FisheryAI is a lightweight Deep Learning system designed to analyze time-series data related to global fisheries and aquaculture. It uses an LSTM-based Data-to-Text architecture to generate context-aware status reports in natural language (Turkish). Optimized for Edge AI using TensorFlow Lite, capable of running on low-resource devices like Raspberry Pi.

Tech Stack

PythonTensorFlowTFLiteDockerLSTMNLP

Key Features

01Offline Intelligence - No cloud dependency, ~0.3ms inference time
02Context-aware Data-to-Text generation
038-bit quantized TFLite model (~108 KB)
04Batch mode for scripting and automation
05Docker containerization ready