Yang Longwei contributed to both the bytedance/deer-flow and vxcontrol/pentagi repositories by expanding AI integration options and enhancing configuration management. Over two months, Yang implemented Novita AI as an optional LLM provider, enabling seamless switching between AI backends and laying the foundation for future vendor diversification. The work involved updating YAML configuration files, refining Docker-based deployment processes, and providing clear documentation to streamline onboarding and maintenance. By updating Novita AI models and ensuring provider-agnostic architecture, Yang improved the flexibility and maintainability of backend systems. The engineering approach emphasized clarity, modularity, and ease of adoption for users and developers alike.