Phodal | 全栈工程师

Original link: http://www.phodal.com/blog/prompt-strategy-code-semantic-search/ In the past week, in order to better build the AI ​​Agent framework Chocolate Factory (hereinafter referred to as CF), we added a new application: the code base AI assistant. This article is reproduced from: http://www.phodal.com/blog/prompt-strategy-code-semantic-search/ This site is only for collection, and the copyright belongs to the original author.

Original link: http://www.phodal.com/blog/agent-framework-chocolate-factory/ In the past period of time, we have tried to refining and summarizing our previous AIGC application experience. From the initial ClickPrompt (https://ift.tt/0qRClAs) and ChatFlow, to a series of applications built under our AI + software development organization Unit Mesh: This article is reproduced from: http://www.phodal.com/blog/agent-framework-chocolate-factory/ This site is only for collection,

Original link: http://www.phodal.com/blog/co-unit-exploring-llm-as-team-api/ In the article “LLM as Co-integrator: Reshape Inter-Team Interaction, Continuously Improve Information Alignment”, we said that in order to better use AIGC to improve performance, our second stage should be: let LLM do some collaborative work , such as: building multi-scenario knowledge questions and answers, reducing knowledge retrieval costs, designing team APIs,

Original link: http://www.phodal.com/blog/autodev-1-1-customize-everything/ In the past half a month, we have added more powerful customization capabilities to the open source auxiliary programming tool AutoDev, now you can: This article is transferred from: http://www.phodal.com/blog/autodev-1-1-customize-everything/ This site is only for collection, and the copyright belongs to the original author.

Original link: http://www.phodal.com/blog/llm-as-co-integrator/ At the QCon conference in May, we shared “Exploring New Processes in Software Development: LLM Empowers R&D Efficiency Improvement”. In that sharing, we emphasized that the team does not spend all their time on SDLC, and may only spend 30% to 50% of their time on developing software, and may even only

Original link: http://www.phodal.com/blog/how-to-encoding-prompt/ From April until now, we have continued to write a series of functions for AutoDev. Despite being in development for over three months, we’ve been thinking about and refactoring the way we manage prompts. This article is transferred from: http://www.phodal.com/blog/how-to-encoding-prompt/ This site is only for collection, and the copyright belongs to the

Original link: http://www.phodal.com/blog/autodev-spec-driven-aigc/ A few months ago, we were heading towards the goal of exploring: ** How to combine AIGC’s R&D efficiency improvement? ** Open source AutoDev, as introduced by GitHub: This article is transferred from: http://www.phodal.com/blog/autodev-spec-driven-aigc/ This site is only for collection, and the copyright belongs to the original author.

Original link:http://www.phodal.com/blog/llm-empowered-devops-toolchain/ In the first two articles of this series “[LLM-enabled R&D Effectiveness](https://ift.tt/MCaLEqY”, “[LLM-enabled software architecture](https://ift.tt/MCaLEqY”, ://https://ift.tt/Yhjm2a8″, introduced a series of experiments that we combined BizDevOps and software architecture in LLM. Focusing on these two types of exploration, we built two open source LLM tools: AI-assisted programming tool AutoDev (IDEA plug-in), and architecture governance tool

Original link: http://www.phodal.com/blog/llm-driven-classic-software-development/ With the application of large language models in software development more and more widely, traditional software engineering practices began to be re-focused and mentioned. In areas such as writing clear documentation, doing code reviews and unit testing, we can see that LLM (Large Language Model) can bring a lot of improvement. In

Original link: http://www.phodal.com/blog/opensoure-aigc-book-llm-with-software/ TL;DR version; This article is transferred from: http://www.phodal.com/blog/opensoure-aigc-book-llm-with-software/ This site is only for collection, and the copyright belongs to the original author.

Original link: http://www.phodal.com/blog/llm-reference-architecture/ With the further exploration of the application of LLM in the field of architecture, and the development of ArchGuard Co-mate into the in-depth area, we found more and more common patterns. This article is transferred from: http://www.phodal.com/blog/llm-reference-architecture/ This site is only for collection, and the copyright belongs to the original author.

Original link: http://www.phodal.com/blog/language-api-llm-first-api/ In the past two or three months, when developing software based on large language models, I have been thinking about how to design a large model-first architecture. And as more and more teams join this competition, we will find that the software architecture based on the large language model is different from

Original link: http://www.phodal.com/blog/arch-spec-as-governance-function/ When we designed the architecture of the architecture governance platform ArchGuard 2.0, the point we have been emphasizing is: **tooling based on norms + patterns**. To put it simply, the specification is the synchronicity of the architecture design, and it is also the explicitness of the architecture knowledge. Therefore, when designing an