Edge Ready Time-series ML pipeline
==================

Here, I have implemented a system includes synthetic data generation along with data cleaning, normalization, and outlier removal to ensure high-quality inputs. It incorporates feature engineering techniques such as rolling statistics and FFT, and applies both classical machine learning models and neural networks, including 1D-CNN and LSTM. The solution supports edge deployment using TFLite and ONNX with model quantization for lightweight performance, provides visualization of data and model predictions, and optionally integrates a FastAPI layer for hybrid edge–cloud simulation.

[code](https://github.com/SalmaKazemiRashed/Edge-Ready-Time-series-ML-pipeline/)