📄 中文摘要
大型语言模型（LLM）为各种应用程序的开发带来了新的机遇，包括助手、协作工具和自主代理。然而，许多团队在使用 LLM 时面临同样的问题：在演示中应用程序运行良好，但在生产环境中却不可靠。这是因为单靠提示工程无法满足生产需求。为构建可靠的 AI 系统，需要更强大的方法——流工程。流工程不仅解决了提示工程的局限性，还提供了实际的架构和可实施的示例，帮助工程师在实际应用中提升 LLM 系统的稳定性和可靠性。
提示工程不足：流工程进入生产 LLM 系统
📄 English Summary
Prompt Engineering Is Not Enough: Enter Flow Engineering for Production LLM Systems
Large Language Models (LLMs) have opened up new opportunities for various applications, including copilots, assistants, and autonomous agents. However, many teams encounter the same issue: their applications perform well in demos but become unreliable in production. This is because prompt engineering alone is insufficient for production needs. To build reliable AI systems, a more powerful approach is required: Flow Engineering. Flow Engineering addresses the limitations of prompt engineering and offers practical architectures and implementable examples to help engineers enhance the stability and reliability of LLM systems in real-world applications.