THE AI FACTORY STACK

Based on the diagram above, here are the 7 subjects (layers of The AI Factory Stack) and short explanations for each:

1. LLM = Brain

Explanation: The Large Language Model serves as the brain. It understands requests, reasons through instructions, writes outputs, and maintains context during the interaction.

2. RAG = Research Assistant

Explanation: This component acts as a research assistant, finding the most relevant and up-to-date information needed to answer queries accurately and contextually.

3. VECTOR DB = Knowledge Database

Explanation: This is the knowledge repository. It stores information and enables semantic retrieval, meaning it finds content based on its meaning, not just exact keywords.

4. AI AGENT = Worker

Explanation: The AI Agent is the worker responsible for executing tasks. It decides the next steps, utilizes necessary tools (like accessing data or models), and completes workflows.

5. MCP = Universal Tool Adapter

Explanation: The MCP acts as a universal adapter, allowing the system to connect with various models, apps, files, and databases through one unified interface.

6. GUARDRAILS = Safety Rules

Explanation: These are the critical safety measures. They define the boundaries of what the system can and cannot do, protecting users, data, and actions.

7. EVALS = Testers & Quality Control

Explanation: This layer performs the quality checks. It assesses correctness, safety, usefulness, performance, cost, and reliability of the entire system's operation.