Skip to content

Benrishty/ai-engineering-wiki

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AI Engineering Wiki

A structured, interlinked knowledge base for AI/ML engineering — a "second brain" that compounds: knowledge is ingested once, integrated, cross-referenced, and kept current rather than re-derived on every query.

It's an Obsidian vault of plain markdown. Currently ~52 sources distilled into ~194 interlinked pages.


How it's organized

Three layers:

Layer What it is Read it for
wiki/ The curated knowledge layer. Start here. Distilled, cross-linked understanding
index.md A catalog of every page, one line each Finding what exists before searching
raw/ Immutable primary sources (articles, transcripts, docs) Precise quotes / original detail

Inside wiki/:

  • sources/ — one page per ingested source (TL;DR + key claims).
  • entities/ — people, orgs, tools, models, libraries.
  • concepts/ — techniques, patterns, methods, architectures.
  • topics/ — broad subject areas that connect multiple pages.
  • syntheses/ — cross-source comparisons and analysis.

Pages link to each other with [[wikilinks]]. CLAUDE.md is the agent schema that governs how the wiki is maintained — not reading material.


How to browse it

In Obsidian (recommended): clone the repo and Open folder as vault. You get backlinks, full-text search, and the graph view — the fastest way to explore how ideas connect.

git clone https://github.com/Benrishty/ai-engineering-wiki.git

On GitHub: open index.md and click through the [[wikilinks]].


How to chat with it (Claude Project)

Chat with the whole wiki from your browser or phone — no need to clone anything — by pointing a Claude Project at this public repo.

Setup (~3 minutes):

  1. Go to claude.aiProjectsCreate project. Name it e.g. AI Engineering Wiki.
  2. In Custom instructions, paste the block below.
  3. Add knowledge → connect GitHub → select Benrishty/ai-engineering-wiki. (Optional: scope it to wiki/ + index.md to skip the large raw/ source layer.)
  4. Re-sync the GitHub source whenever the wiki updates so chat stays current.

Sharing note: On Free/Pro plans, Projects are personal — your colleague can simply create their own Project against this same public repo using the instructions below. On Team/Enterprise plans you can share one Project directly.

Custom instructions to paste:

This project is a structured knowledge base ("AI Engineering Wiki") for AI/ML
engineering, maintained as an Obsidian vault of interlinked markdown.

Layers:
- wiki/  — the curated knowledge layer. Prioritize this. Subfolders: sources/
  (one page per ingested source), entities/ (people, orgs, tools, models),
  concepts/ (techniques, patterns), topics/ (subject areas), syntheses/
  (cross-source analysis).
- index.md — the catalog of every page. Consult it first to find what exists.
- raw/   — immutable primary sources. Use only when wiki/ lacks detail; quote precisely.

When answering:
- Cite the wiki pages you used by name, e.g. [[mixture-of-experts]]. Pages
  cross-link via [[wikilinks]] — follow them to find related material.
- Prefer wiki/ summaries; fall back to raw/ only for precision.
- If something isn't in the wiki, say so rather than inventing it.
- Be concise and encyclopedic: definition first, then supporting detail.

Zero-setup alternative: because this repo is public, you can also paste its URL into DeepWiki to get an instant, shareable chat-with-the-repo link — no account needed on the other end.

About

Personal LLM Wiki for AI/ML engineering. Incrementally maintained markdown knowledge base built per the Karpathy LLM Wiki pattern; schema in CLAUDE.md.

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages