# Project 9: Personal Learning OS

## What We're Building

A personal AI operating system that orchestrates memory (Project 5), retrieval (Projects 3–4), agents (Project 8), and evaluation (Project 7) into a unified system for personal knowledge management and learning. This is the direct prototype for StarcallOS architectural patterns.

## Why We're Building It

This is the synthesis project. Every previous project taught you a component. Now you design the system that makes them work together. Personal AI operating systems are the emerging category — the AI that knows you, learns from you, and works for you across time.

## Learning Objectives

- [ ] Design a multi-component AI system with clear interfaces between components
- [ ] Integrate memory, retrieval, agents, and evaluation into one product
- [ ] Design for extensibility: how do you add a new component without breaking others?
- [ ] Define the data model for a personal knowledge graph
- [ ] Implement a query router that chooses the right component for each task
- [ ] Evaluate the system holistically (not just individual components)
- [ ] Apply the "Build vs. Buy" lens to a full system architecture

## Key Concepts

System design, orchestration, query routing, knowledge graphs, data models, component interfaces, extensible architecture, AI product thinking

## Core Engineering Problem

**Problem:** You have memory, search, agents, and evaluation working independently. How do you compose them into a coherent system that a user can actually rely on — without the complexity growing unmanageable?

## Time Estimate

**Total:** 15–20 hours

## Startup Lens

This project IS a startup idea. Personal AI OS is the product category that Apple, Google, Microsoft, and dozens of startups are racing to define. What you build here is a working proof-of-concept for a real product.

## Key Files

```
code/
  os_app.py        — User-facing REPL and one-shot CLI
  learning_os.py   — Main orchestration layer
  router.py        — Route requests to SAVE, RECALL, TASK, or CHAT
  knowledge.py     — Personal knowledge graph data model
  subsystems.py    — Provided stand-ins for memory, retrieval, agent, and chat workers
  evaluate_os.py   — Holistic router/system evaluation
  config.py        — Provider and runtime configuration
  tests/           — Offline guiding tests for each learner-owned milestone
  requirements.txt
  README.md
```