Radio modulation recognition with CNN, CLDNN, CGDNN and MCTransformer architectures. Best results were achieved with the CGDNN architecture, which has roughly 50,000 parameters, and the final model has a memory footprint of 636kB. More details can be found in my bachelor thesis linked in the readme file.

Code repository for 'Machine Learning based framework for Drone Detection and Identification using RF signals'

SpectroSense is an open-source RF signal analysis platform that combines artificial intelligence with traditional signal processing techniques to automatically identify and classify radio frequency signals from spectrograms.

1st Place Winner - EDTH Switzerland 2025 - A distributed system for triangulating RF signal sources using multiple sensors and edge computing hubs.

Production-ready FastAPI service and PyTorch pipeline for RF Signal Modulation classification. Includes 1D-CNN training, early stopping validation, ONNX Runtime inference engine, and dynamic experiment metrics and artifacts tracking via MLflow proxy.

A deep learning system for RF modulation recognition using PyTorch and MATLAB. Classifies raw I/Q signals and analyzes the sim-to-real domain gap on the RadioML dataset via an interactive Streamlit dashboard.