Laila Belakhdar

Troubleshooting video streaming problems: An experimental approach backed by Machine Learning Contributed to a machine learning research project at Inria focused on diagnosing video streaming slowness. Built a complete testbed to diagnose video streaming issues using Mininet-WiFi, Docker, and Selenium. Developed a custom DASH.js plugin to extract real-time QoE metrics, and collected data via a FastAPI backend stored in MongoDB. Applied the ITU-T P.1203 model for quality estimation. Trained machine learning models to detect anomalies and identify root causes. Automated the full workflow for reproducible testing under varied network conditions. Visualized QoS/QoE data using Grafana and Python libraries.