Mobile Traffic Forecasting, Anomaly Detection & Signal Reconstruction

Telecom capacity and performance planning requires understanding future traffic behavior while also distinguishing genuine degradation from normal temporal variation.

Built deep-learning pipelines for 4G/5G traffic forecasting using LSTM, Seq2Seq, CNN-based and autoencoder approaches.

Applied Isolation Forest and SVM-based techniques to historical OSS/KPI data to identify anomalous network behavior.

Used REST APIs as part of the processing and serving workflow.