Home / AI Arena / Building AI Agents
Building AI Agents
A hands-on series that walks through building LLM-powered agents from first principles — starting with a single LLM call and building up through retrieval-augmented generation, autonomous agent loops, multi-agent coordination, and production safety. Examples are implemented with LangChain and LangGraph running against a local model.
-
Part 1
Introduction to AI Agents
What AI agents are, why they exist, and how to set up a local environment to start building them with LangChain and LangGraph.
-
Part 2
LLM Clients and Chains
How to talk to LLMs from code — ChatOpenAI vs OpenAI, composing chains with LCEL, and the invoke, batch, and stream invocation patterns.
-
Part 3
Chunking Strategies
How to break documents into retrieval-friendly chunks — fixed-size, structure-aware, language-aware splitting, and choosing the right chunk size.
-
Part 4
Embeddings and Vector Databases
Turning document chunks into vectors, storing them in pgvector, and running similarity search to find relevant content at query time.
-
Part 5
RAG Strategies
Retrieval-Augmented Generation from the naive baseline through RAG Fusion, RAPTOR, GraphRAG, agentic RAG, and self-evaluating pipelines.
-
Part 6
Query Transformation
Making retrieval smarter before it hits the vector database — query rewriting, multi-query generation, HyDE, and logical routing.
-
Part 7
Agent Architectures
Building autonomous agents with LangGraph — the ReAct loop, reflection, supervisor multi-agent systems, subgraphs, dynamic tool selection, and human-in-the-loop.
-
Part 8
Agent Guardrails
How to validate and filter agent inputs and outputs — classifier nodes, structured output validation, prompt injection defence, and safe fallback responses.