# MumbleKey Blog > Fred Chu on AI agent engineering, one-person-company automation, and cycle-based trading. > Bilingual: Traditional Chinese (zh-TW, primary) and English. Paired posts share a slug: > https://blog.mumblekey.com/zh/posts// and https://blog.mumblekey.com/en/posts//. > Every post has a plain-Markdown version at the same URL with a .md suffix (drop the trailing slash). ## Posts - [CLAUDE.md Is a Score: Running a One-Person Company as an AI Conductor](https://blog.mumblekey.com/en/posts/claude-md-as-score.md): I don't have employees, and I don't pay for a pile of SaaS. I restructured my work into a system of AI agents: two Macs, one rule file that keeps growing, and four agent roles that each stay in their lane. Here's the architecture and the traps I hit. - [From 3 Hours to 30 Minutes: My Automated Chinese Subtitle Pipeline](https://blog.mumblekey.com/en/posts/subtitle-pipeline-3h-to-30min.md): A subtitle production line running on Apple Silicon — local ASR, on-screen term extraction, LLM correction, and a fail-safe quality gate. This post covers the architecture, the term-learning loop, and three lessons that took it from "it runs" to "it's reliable." - [CLAUDE.md 是樂譜:一人公司的 AI 指揮家模式](https://blog.mumblekey.com/zh/posts/claude-md-as-score.md): 我沒有請員工,也沒有訂閱一堆 SaaS。我把工作重組成一套 AI agent 系統:兩台 Mac、一份會長大的規則檔、四個各司其職的 agent 角色。這篇講架構跟踩過的坑。 - [從 3 小時到 30 分鐘:我的中文字幕自動化 pipeline](https://blog.mumblekey.com/zh/posts/subtitle-pipeline-3h-to-30min.md): 一條跑在 Apple Silicon 上的字幕生產線:本地 ASR、螢幕畫面術語抽取、LLM 校正、防呆品質 gate。這篇講架構、術語學習迴圈,以及三個讓它從「能跑」變「可靠」的教訓。 ## Demos - [Subtitle Pipeline Visualized: One Line's Six-Stage Journey](https://blog.mumblekey.com/en/demos/subtitle-pipeline-flow/): Walk through an automated subtitle production line: watch a real subtitle line evolve from raw ASR homophone errors into timed, quality-gated output. - [字幕 pipeline 流程視覺化:一條字幕的六段旅程](https://blog.mumblekey.com/zh/demos/subtitle-pipeline-flow/): 互動走一遍字幕自動化生產線:看一條真實字幕行從 ASR 同音字錯誤,一步步變成帶時間軸的成品。 ## Feeds - [RSS zh-TW](https://blog.mumblekey.com/zh/rss.xml): full-content feed, Traditional Chinese - [RSS en](https://blog.mumblekey.com/en/rss.xml): full-content feed, English - [Full content](https://blog.mumblekey.com/llms-full.txt): all posts concatenated as Markdown