Independently maintained a Minecraft server for several years, with an average of 200+ daily active players, handling multiple major version updates and server migrations. Mitigated malicious stress tests using reverse proxy and load balancing with Bungeecord, and defended against DDoS attacks with elastic IPs and other cloud services.
Developed several Bukkit-based plugins in Java, implementing SQL injection prevention, asynchronous I/O operations to prevent main thread blocking, and adding various game features such as title systems, guild systems, currency systems, shop systems, GUI interfaces, group chat bots, etc., and configured and maintained numerous third-party plugins to enhance gameplay.
Designed and implemented a database schema with 24 tables using MySQL, optimized indexes, normalized to fourth normal form (4NF), and mapped the database using Django ORM. Developed RESTful APIs using the Django framework, documented with Swagger, performed API testing with Postman, and managed frontend-backend integration.
Deployed the production environment using uWSGI and Nginx, and containerized the application with Docker.
Implemented features such as session management, email service, image object storage, posting and commenting, online veterinary consultation, dog walking route tracking, and BERT-based comment sentiment analysis.
Built a weighted directed graph model, calculated similarity between artists and music genres recursively using graph DFS and improved cosine similarity algorithms, and introduced a decay factor to reflect the diminishing effects of indirect influence, summarizing the topological network of musical influences.
Cleaned and preprocessed the dataset, discovered and reported a containment error in the official dataset during recursion. Evaluated the dispersion and weight variation of musical features using entropy, reflecting dynamic changes, and used Pearson correlation coefficient to identify the most influential features of leading musicians.
IoT device control app built with Swift and SwiftUI. Communicated with backend APIs via HTTP requests for real-time information and control of IoT devices