1. Participated in the analysis and mining of massive data to build multiple user models in the field of hydropower. Applied Python as the main programming language. 2. Designed and implemented a real-time anomaly detection model based on Variational Auto-Encoders. Conducted experiments, and summarized the tuning method for the input window of the model, which can make the F2 value of the anomaly detection model reach more than 90%. 3. Accelerated the distributed implementation of existing algorithms and models, which process the same data nearly 25% faster than before, enriching the common parallel algorithm library in the company. 4. Implemented a complete reliability assessment system for hydroelectric power plant discharge pumps and deployed it in the production environment.