Zhihao Yang developed the Orbit Angle Prediction Module for the Julek-AK/AE2224-I-B04 repository, establishing a data-driven baseline for orbit angle forecasting using Python and PyTorch. He designed and implemented a neural network with multiple hidden layers and ReLU activation, integrating a training loop with SGD optimization and MSE loss to support rapid experimentation. In addition to building the initial model and data-loading pipeline, Zhihao refactored legacy LSTM prediction logic to prepare for future predictor migration, aligning data interfaces for maintainability. His work demonstrated depth in data preprocessing, deep learning, and machine learning, focusing on production-readiness and technical debt reduction.