AI Application Development Engineer | AI travel planning assistant with LLM prompts/workflow orchestration (Coze, FastAPI, LangChain)
Worked on building LLM-based chat and assistant features that required designing prompts, orchestrating multi-turn dialogue, and shaping structured outputs for downstream use. Implemented streaming, exception handling, logging, and tool-integrated agent reasoning to ensure reliable generation behavior. Integrated Coze workflows to generate travel plans in a structured way aligned with application needs.• Developed a DeepSeek-style streaming dialogue AI chat application using LLM APIs and backend services.• Implemented an idiom chain automation using LangChain as part of prompt/workflow-driven generation.• Built ReAct agent reasoning workflows with LangChain and integrated external weather/travel tools.• Added data cleaning and robust logging around generated outputs and workflow execution.