Awesome Papers with Claude Code

A curated collection of AI/ML papers turned into interactive learning courses powered by Claude Code.

Part of the Papers with Claude Code project — see ainblockchain/papers-with-claudecode for the full platform (frontend, blockchain integration, course builder, Cogito Node).

How it works

Each paper directory contains:

• paper.json — Paper metadata (title, authors, arXiv ID, GitHub repo)
• Course subdirectories — Each is a self-contained learning path with a CLAUDE.md tutor prompt, knowledge graph, and lesson content

Open any course directory in Claude Code and start chatting. The AI tutor will guide you through the paper with explanations, analogies, code snippets, and quizzes.

Papers

Paper	Authors	Year	Courses
Attention Is All You Need	Vaswani et al. (Google Brain)	2017	10
Adam: A Method for Stochastic Optimization	Kingma & Ba (Google Brain / OpenAI)	2014	1
An Image is Worth 16x16 Words (ViT)	Dosovitskiy et al. (Google Research)	2020	1
Direct Preference Optimization (DPO)	Rafailov et al. (Stanford)	2023	1
T5: Text-to-Text Transfer Transformer	Raffel et al. (Google)	2019	1
HippoRAG	Gutiérrez et al. (Ohio State)	2024	1
LLaMA 2	Touvron et al. (Meta AI)	2023	1
YOLOv3	Redmon & Farhadi (U. Washington)	2018	1

Quick start

# Clone and enter a course
git clone https://github.com/ainblockchain/awesome-papers-with-claude-code.git
cd awesome-papers-with-claude-code/attention-is-all-you-need/bible

# Open in Claude Code and start learning
claude

Then just chat: "teach me about self-attention", "show the graph", "next", "exercise".

Course structure

paper-name/
├── paper.json              # arXiv metadata
├── course-variant/
│   ├── CLAUDE.md           # AI tutor instructions
│   ├── README.md           # Course overview
│   ├── knowledge/
│   │   ├── graph.json      # Knowledge graph (nodes + edges)
│   │   └── courses.json    # Lessons and quizzes
│   └── .learner/           # Progress tracking (created on first use)
│       ├── profile.json
│       └── progress.json
└── another-variant/
    └── ...

Learner features

• Knowledge graph — Visualized as a Mermaid diagram showing concept dependencies
• Progress tracking — Completed concepts marked, next concept recommended via graph topology
• Friends — Share progress via git branches, see friends' positions on the graph
• Quizzes — Each concept ends with a quiz (multiple choice, predict-the-output, true/false)

On-chain integration

Courses can optionally record progress on the AIN blockchain knowledge graph:

• Each concept completion → ain.knowledge.explore() on-chain
• Community frontier map → who explored what, to what depth
• x402 micropayments → premium content gated by AIN-token payments
• Visualize on AINscan

See ainblockchain-integration for details.

Contributing

1. Find a paper on arXiv with open-source code
2. Use the Course Builder or generate manually with Claude Code
3. Submit a PR with paper-slug/paper.json + one or more course directories

Related

• paperswithclaudecode.com — Full platform with explore, learn, village, and course builder
• ainblockchain/papers-with-claudecode — Platform source code
• ainblockchain/ain-js — AIN blockchain SDK with knowledge module
• AINscan — Blockchain explorer with knowledge graph visualization

License

MIT