AI Decision Graph
Concepts

Decision nodes, not dictionary entries.

Each page answers when to use the concept, when to avoid it, and what to learn next.

RAG and retrieval
rag medium

Chunking

Chunking splits source material into retrievable pieces for search and RAG.
rag medium

Embedding

Embeddings turn text or other data into vectors so similar items can be compared mathematically.
rag high

Fine-tuning

Fine-tuning adapts a model's behavior using examples, usually to improve style, format, or narrow task performance.
rag medium

Hybrid Search

Hybrid search combines semantic vector retrieval with keyword or full-text search.
rag low

Long Context

Long context means putting a large amount of information directly into the model's prompt window.
rag medium

RAG

Retrieval-Augmented Generation lets an LLM retrieve external knowledge before answering.
rag medium

Reranking

Reranking reorders retrieved candidates so the most relevant evidence is placed first.
rag medium

Vector Database

A vector database stores embeddings and retrieves nearby vectors efficiently.
Agents and workflows
agents high

Agent

An AI agent uses a model to choose steps, call tools, observe results, and continue toward a goal.
agents low

Automation

Automation executes a known process with predefined triggers, steps, and outcomes.
agents medium

Function Calling

Function calling is a model interface pattern where the model emits structured arguments for predefined functions.
agents medium

MCP

Model Context Protocol is a standard way for AI applications to connect with external tools, data, and context servers.
agents medium

Tool Calling

Tool calling lets a model request external actions such as search, database lookup, calculation, or API calls.
agents medium

Workflow

A workflow is a controlled sequence of steps where the application decides the path.
Product engineering
engineering low

Prompt Engineering

Prompt engineering shapes model behavior through instructions, examples, context, constraints, and output rules.
engineering low

Structured Output

Structured output asks a model to return data in a predictable schema instead of free-form prose.
Evaluation and safety
safety medium

Agent Evaluation

Agent evaluation measures whether an agent chooses useful steps, calls tools safely, and reaches the goal under realistic conditions.
safety medium

Evaluation

Evaluation measures whether an AI system behaves correctly on realistic examples.
safety medium

Guardrails

Guardrails constrain AI behavior through validation, policies, permissions, fallbacks, and human review.
safety medium

Tool Permission

Tool permission defines what an AI system is allowed to call, with what arguments, and under which user or system authority.