● Optimized real-time data sampling using coreset selection algorithms to achieve 50-60% training data efficiency, maintaining model performance while minimizing computational overhead in streaming environments. ● Developed a carbon footprint library for edge devices to monitor power consumption and calculate CO2 emissions based on device-specific metrics. ● Implemented advanced continual learning strategies using the Avalanche framework, including dataset preparation, model selection, edge integration, and hyperparameter tuning for reproducible industrial applications.