During my internship at VGU's ECE Lab, I took on the formidable challenge of developing a time-series prediction model for air quality data sourced from legacy.ideasky.app. The scarcity of pre-trained models, combined with the challenge of black-out data, presented significant hurdles. While direct approaches to handling such data are rare, my innovative solution proved highly successful. By integrating elements of Temporal Convolutional Networks (TCN) and Long Short-Term Memory (LSTM), I devised a hybrid model capable of directly addressing black-out datasets. This achievement, while not widely attempted, marked a significant success in my internship experience.