Agent runs Agent 运行
from experiments to repeatable workflows 从实验到可复用流程AI Agent / Workflow / Evidence
Hi, i'm mumong
Build useful AI workflows. 把 AI Agent 落到真实工作流里。
Input Real Needs
Output Reusable Flow
From real needs to reusable AI workflows. 拆解业务流程，连接工具、数据与知识库，让 AI 从 Demo 变成可复用的生产力。
Contact Me 联系合作Selected Experiments
Real experiments from AI workflow practice 一些真实的 AI 工作流实践tokens Token
processed through real task work 用于真实任务拆解与生成workflows 工作流
AIOps, observability, and content systems AIOps / 可观测性 / 内容生产notes 笔记
turned into public notes and methods 沉淀为公开文章和方法About
About Me
我 关 注 A I A g e n t 如 何 进 入 真 实 工 作 流 。 从 需 求 拆 解 、 上 下 文 工 程 、 工 具 调 用 到 结 构 化 输 出 ， 让 A I 不 只 回 答 问 题 ， 而 是 完 成 任 务 、 沉 淀 流 程 、 提 升 效 率 。
I c a r e a b o u t h o w A I A g e n t s e n t e r r e a l w o r k f l o w s . F r o m n e e d a n a l y s i s a n d c o n t e x t e n g i n e e r i n g t o t o o l u s e a n d s t r u c t u r e d o u t p u t , t h e g o a l i s n o t b e t t e r a n s w e r s , b u t r e p e a t a b l e w o r k .
Capabilities
Workflow Design
Workflow Decomposition 业务流程拆解
Start from a real need, map repeated actions, decision points, information paths, and deliverables, then turn complex work into executable Agent flows. 从真实需求出发，识别重复动作、判断节点、信息路径和最终交付物，把复杂任务拆成 Agent 可以执行的流程。
Agent Building
Agent Workflow Building Agent 工作流构建
Design agent boundaries, tool paths, execution steps, handoff mechanisms, and structured outputs so agents can complete specific work reliably. 设计 Agent 的角色边界、工具路径、执行步骤、Handoff 机制和结构化输出，让 Agent 能稳定完成具体任务。
Context Engineering
Context Engineering 上下文工程
Organize metrics, logs, events, runbooks, historical cases, and tool results around the task, so analysis is evidence-based. 围绕任务目标组织指标、日志、事件、Runbooks、历史案例和工具结果，让 Agent 基于证据分析。
Observability
Observability Base 可观测性底座
Connect metrics, logging, tracing, and tool calls to provide a stable data base for agents working with real systems. 串联 Metrics、Logging、Tracing 与工具调用，为 Agent 分析真实系统提供稳定的数据基础。
AI Productivity
AI Productivity AI 工具提效
Use Codex and agent tools to reshape everyday work, turning slides, video, websites, resumes, and documents into reusable workflows. 使用 Codex 和 Agent 工具改造日常工作流，将 PPT、视频、网页、简历和材料生成变成可复用流程。
Projects
Real Workflows
AI Agent / AIOps
GitHub AIOps Agent
Analyze Pod incidents and produce structured operations reports. 自动分析 Pod 异常，生成结构化运维文档。
For Kubernetes Pod incidents, the agent calls tools for status, logs, events, and metrics, then combines runbooks into an evidence chain. 面向 Kubernetes Pod 异常场景，Agent 自动调用工具获取状态、日志、事件和指标，结合 Runbooks 构建证据链，帮助运维人员快速缩小问题范围。
Cloud Native / Observability
GitHub Observability Base
Connect metrics, logging, and tracing into one evidence system. 把 Metrics、Logging、Tracing 串成统一可观测底座。
Deploy observability components with Helm and connect collection, storage, query, and visualization for real AIOps data. 将可观测性组件通过 Helm 统一部署，打通采集、存储、查询和可视化链路，为 AIOps Agent 提供真实系统数据基础。
HTML PPT / Multi-Agent
Read Article Hermes PPT Automation
Turn PPT creation into a reviewable, repeatable agent workflow. 把 PPT 创作拆成可验收、可返工的 Agent 流程。
For HTML slide decks, Hermes splits intake, spec, research, design, review, rework, and package into explicit agent tasks with review gates. 围绕 HTML PPT 制作，把需求、规格、调研、设计、评审、返工、交付拆成明确任务，让速度来自生成器，确定性来自质量门控。
Articles
Field Notes 文章笔记
Writing is where experiments become reusable methods. 把实践、踩坑和方法写下来，让一次实验变成可复用的经验。
AI Agents ai-agent
Creating a High-Quality PPT Video with Hermes and PPT Skill 用 Hermes 和 PPT Skill 制作一条高质量 PPT 视频
A real workflow note on using Hermes AI Agent, profiles, skills, and orchestration to produce a PPT-based video, with the full recording available on Bilibili. 记录我如何用 Hermes AI Agent、profile、skill 和工作流来完成 PPT 内容生产，并把过程录制成一条可复用的视频案例。
Jun 16, 2026 2026/06/16 LLM llmProduction Deployment for Local LLMs: vLLM, Ray, Kubernetes, and VRAM Strategy 本地大模型生产化部署：vLLM、Ray、Kubernetes 与显存策略
A production-focused guide to distributed vLLM deployment with Ray, Kubernetes orchestration, tensor and pipeline parallelism, VRAM planning, and operations practices. 围绕 vLLM 与 Ray 的分布式容器化部署，整理 Kubernetes 编排、Ray Serve、张量并行、流水线并行、VRAM 分配和生产运维策略。
Jun 5, 2026 2026/06/05 LLM llmLocal LLM Deployment and Inference Optimization: From Ollama to vLLM 本地大模型部署与调优：从 Ollama 到 vLLM
An overview of local LLM inference bottlenecks, quantization formats, Ollama's role and limits, and vLLM optimizations such as PagedAttention and continuous batching. 梳理本地大模型推理的核心瓶颈、量化格式选择、Ollama 的定位与限制，以及 vLLM、PagedAttention、连续批处理带来的吞吐优化。
Jun 5, 2026 2026/06/05 AI Agents ai-agentDeploying and Evaluating an AIOps Agent AIOps 智能运维 Agent 部署、评估与小模型测试
Deployment notes, usage flow, feature validation, MTTR, root-cause accuracy, evidence completeness, and small-model testing for an AIOps Agent. 记录智能运维助手的 Kubernetes 部署方式、使用流程、核心功能验证、MTTR、根因准确率、证据完整率和小模型测试结果。
Jun 5, 2026 2026/06/05 AI Agents ai-agentLangGraph and MCP Orchestration for an AIOps Agent AIOps 智能运维 Agent 的 LangGraph 与 MCP 编排设计
A deep dive into API design, configuration, LangGraph state machines, federated multi-cluster Agents, scheduling, and MCP tool decoupling. 拆解智能运维助手的接口层、配置层、LangGraph 状态机、多集群联邦 Agent、调度引擎和 MCP 工具解耦方案。
Jun 5, 2026 2026/06/05 AI Agents ai-agentAIOps Agent Architecture Design AIOps 智能运维 Agent 总体架构设计
A structured overview of an AIOps Agent, covering requirements, system architecture, streaming output, context management, automated remediation, and observability. 梳理智能运维助手 3.0 的需求分层、总体架构、技术路线、流式输出、上下文管理、自动修复和可观测性设计。
Jun 5, 2026 2026/06/05Resume
Resume 简历
A compact online view of experience, projects, education, and public contact paths. 在线查看完整经历、项目、教育和公开联系方式。
Contact
Let's Build Useful AI Workflows
Open to conversations around AIOps, agent workflows, content automation, and productivity tools. 欢迎交流 AIOps、Agent 工作流、内容自动化和效率工具。