A Chinese-language wiki for breaking into AI product management

What it does AIPM-Wiki is a structured knowledge base, not code: six numbered sections covering a learning roadmap, AI fundamentals (ML, LLMs, prompt engineering, a glossary), core PM skills (PRDs, model evaluation, data labeling, cost estimation), case-study teardowns of products like ChatGPT, Perplexity, and Sora, an interview question bank with 18 experience write-ups, and a resources index. The pitch is that a PM with no AI background can read it front to back and come out with a working framework — and a fighting chance in interviews.

The interesting bit It’s an Obsidian vault with bidirectional links, so courses, interview questions, and case studies cross-reference into a knowledge graph rather than a flat list of files. The author also states most of the content was written and maintained using Obsidian plus the Claudian plugin (in-vault Claude conversations), which is a nice meta touch: an AI-PM wiki built with an AI-assisted workflow.

Key highlights

• Interview bank follows a fixed template: question → what it tests → reference answer → follow-ups → related reading
• Tech and pricing claims are sourced from official Anthropic/OpenAI/Google docs with access dates noted
• Multiple entry points by career stage: career-switcher basics, in-role advancement, and a hands-on “internal IT service assistant” practice track with 12 fictional policies and 30 evaluation items
• Readable three ways: GitHub web, Obsidian vault locally, or a read-only Feishu mirror (synced one-way, may lag)
• CC BY-NC-SA 4.0 licensed — remix with attribution, no commercial use

Caveats

• Chinese-language content throughout; not useful if you can’t read it
• The README itself notes that update dates don’t mean every historical fact is re-verified, and teaching examples aren’t real production results
• Content is opinionated curriculum from one author’s career-switch experience, not a neutral reference

Verdict Worth a star if you’re a Chinese-speaking PM (or aspiring one) prepping for AI product roles — the sourcing discipline alone beats most interview-prep dumps. If you’re an engineer looking for technical depth, this isn’t your repo.