AIPA → Certificates

Certificates

Two credentials anchor the program: Personal AI — build an AI you own in a Codespace — and Sovereign AI — deploy that build to a server you control.

Certificate 1

Personal AI

An AI that the student owns, built and run inside a GitHub Codespace.

This certificate is a standalone outcome: by the end of Chapter 1, you have a working personal AI and a clear path to scale it. You establish the conceptual frame — intelligence as process rather than possession, metabolism of knowledge — and build the stack where it lives: schema, load, query, extend, and integrate a knowledge graph as the metabolic hub of your agent.

The escalation architecture runs through the book: 1B → 3B → 8B → Cloud. You run small local models (1B, 3B), larger local or cloud-backed models (8B), and know when to escalate to cloud. Labs 0–3 deliver the graph pipeline and wire it into the agent.

  • Environment: GitHub Codespaces (devcontainer, zero local setup)
  • Deliverable: Working personal AI at a chosen scale (1B, 3B, 8B, or cloud) and a documented escalation path
  • Core idea: The knowledge graph is the hub; the student owns the code, the data, and the runtime in their Codespace
Certificate outcome. A personal AI you own, running in a Codespace you control, with a clear escalation path and an understanding of intelligence as the metabolism of knowledge.
Certificate 2

Sovereign AI

Deploy the build from Personal AI to a server you own — nominally a Mac Mini M4 (or equivalent).

This certificate completes the arc: you take the agent you built in the Codespace and run it on infrastructure you control from the OS up. You install Linux (or use the host OS) and build the runtime stack from scratch — users, packages, Python, GPU drivers when needed — then deploy your AI locally and decide when, if ever, to escalate to the cloud.

The same 1B → 3B → 8B → Cloud architecture now runs on your own hardware. Themes: local-first for privacy and sovereignty, full control and reproducibility, and the epistemology of who controls the model and the data. Labs build the from-scratch environment, local deployment and benchmarking, and a documented escalation path.

  • Environment: Your own server — e.g. Mac Mini M4 (or bare metal / VM with Linux)
  • Deliverable: The same agent (orchestrator + tools) running on a stack you own from the kernel up; documented setup and escalation policy
  • Core idea: Sovereignty is not only technical — it is a stance on who shapes knowledge flows and who can be held accountable
Certificate outcome. Your AI running on a stack you own: same agent, different infrastructure. You decide when to stay local and when to burst to cloud; governance and compliance are constraints you design.

The two certificates form a single path:

Personal AI (Codespace) → build the agent you own

↓

Sovereign AI (your server) → deploy that build to a machine you control

← Back to AIPA