Turn 10,994 Notes Into Memory - Paul Iusztin, Decoding AI & Louis-François Bouchard, Towards AI

2026-06-26 · Source: AI Engineer · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Advanced, extended

Summary

The AI Research OS, developed by Paul Iusztin and Louis-François Bouchard, is a personalized system designed to manage and utilize extensive research notes, addressing limitations of public LLMs for complex, evolving projects. This system, available as a GitHub repository, integrates personal knowledge bases from platforms like Obsidian, Readwise, NotebookLM, and GitHub. It employs a three-layer architecture: immutable raw files, a simple index.yaml cataloging data and summaries, and a dynamically evolving LLM-generated wiki. The wiki, which can be queried token-efficiently, creates derivatives like comparisons, concepts, and notes, and updates based on user interactions. The system prioritizes file-based storage for inspectability and aims to transform research into actionable work, supporting projects from articles to codebases.

Key takeaway

For AI Engineers building agentic systems or managing extensive research, consider implementing a personalized AI Research OS. This approach overcomes public LLM limitations by integrating your diverse knowledge sources into a dynamic, file-based wiki. You can utilize the provided GitHub repository and Cloud Code plugins to create a token-efficient, evolving knowledge base, ensuring your agents have consistent, personalized context for complex projects and avoiding repetitive context feeding.

Key insights

A personalized, file-based AI research OS dynamically manages and evolves knowledge from diverse sources for complex projects.

Principles

• Personal knowledge bases enhance AI agent context.
• File-based systems simplify personal research OS.
• Dynamic wikis evolve with user interaction.

Method

The system uses a three-layer architecture: raw files, an index.yaml catalog, and an LLM-generated wiki. Agents query the index, then source wiki pages, and finally raw sources for token-efficient information retrieval.

In practice

• Integrate Obsidian, Readwise, GitHub into a unified system.
• Use Cloud Code plugins for deep research and ingestion.
• Scope wikis to specific projects for focused work.

Topics

• AI Research OS
• Agentic AI Engineering
• Knowledge Management
• Context Management
• Obsidian Integration
• LLM Knowledge Bases

Best for: Machine Learning Engineer, AI Scientist, Research Scientist, AI Engineer, Software Engineer, AI Architect

Related on AIssential

• I keynoted alongside OpenAI and DeepMind — A personalized AI research OS transforms personal notes into compounding agent memory via a file-based, LLM-maintained wiki.
• LLM Wiki Revolution: How Andrej Karpathy’s Idea is Changing AI — LLM-powered wikis build persistent, interconnected knowledge at ingestion, overcoming RAG's query-time processing limitations.
• How This Creator Turned 7,000 Obsidian Notes Into A Lucrative Research Engine — Connecting a large personal knowledge base to an AI agent swarm can transform digital clutter into a lucrative, autonomous research engine.
• Build an AI knowledge fabric for your organization — An AI knowledge fabric provides structured, dynamic context for autonomous agents, preventing hallucinations and improving efficiency.
• On Accelerating Grounded Code Development for Research — Lexical search and structured workflows can effectively ground coding agents in dynamic, domain-specific research.
• PersonalAI 2.0: Enhancing knowledge graph traversal/retrieval with planning mechanism for Personalized LLM Agents — PAI-2 enhances LLM factual correctness and reduces hallucinations via dynamic knowledge graph traversal and planning.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Engineer.