Role: Research Intern at the UCLA Cognitive Reconfigurable Embedded Systems Lab.
Signal detection: Implemented K-means models with TensorFlow and Keras for UAV signal detection, spectrum localization, and RF fingerprinting from spectrograms, reaching 99.5% transmission detection precision and 90.9% same-model drone classification accuracy.
Computer vision: Integrated YOLO-based detection with CNNs, OpenCV, and MATLAB to improve wireless signal localization through bounding-box regression, improving prediction accuracy by 60%.
Data pipeline: Used AWS Lambda and S3 to collect and analyze real-time data from network transmitters and receivers, improving data quality and model reliability by 70%.