This project implements an automated fishing AI for Genshin Impact using a combination of deep learning and reinforcement learning. YOLOX, an object detection model, handles fish localization, fish type recognition, and fishing rod positioning through computer vision. A DQN (Deep Q-Network) agent learns to control timing of clicks to land fish within optimal force ranges. The system uses transfer learning and semi-supervised learning for training, and includes traditional image processing components via OpenCV for non-learnable detection tasks.