Agentic AI Mastery: From N8N to DSPy to LangGraph

Who is Rhys? How should you take this course?

Brief intro into what is N8N? What are benefits of N8N, including its open-source nature, flexibility, and visibility.

Explain how n8n's observability features enhance automation maintenance and debugging processes.

Identify N8N components such as workflows, nodes, credentials, executions, evaluations, variables, and insights.

Explain how N8N agents combine workflows, triggers, and AI nodes to create autonomous systems.

Describe the steps to install and run NAN locally using NPX.

Analyze the workflow of a joke-generating agent, including form submission, joke creation, and joke evaluation.

Evaluate the agent's performance based on the average rating from five personas and determine whether to loop back for a new joke or display the final result.

Explain how expressions allow dynamic data manipulation within N8N workflows, referencing external nodes and computing values.

Apply JavaScript operations, date formatting, and AI assistance to troubleshoot common expression-related issues like missing data or incorrect node references.

Describe the process of building a blog writing agent using NAN, including form submission, outline generation, section writing, and final editing.

Explain the steps involved in calculating and displaying the average reading time for the generated blog post.

Analyze competitor URLs to extract and normalize blog content links.

Synthesize summaries of new content using AI and send a digest email.

Analyze the performance of Apple and Microsoft stocks using a model-agnostic agent node.

Design a workflow that retrieves price history and performs financial analysis with code execution capabilities.

Identify the key steps in building an invoice extractor agent using AI and structured output parsing.

Analyze alternative implementation patterns for optimizing the agent's performance and error correction.

Analyze the impact of different prompting techniques on the accuracy of joke classification.

Evaluate the effectiveness of NN's evaluation tools for optimizing AI agent performance in production.

Describe the steps to build a deep research agent using GPT-4o Mini and Open AI search.

Explain how the agent uses tool calling, output parsing, and autofixing to generate a well-structured answer.

Identify the key components of a customer support agent, including chat triggers, JavaScript database simulation, and AI agent configuration.

Analyze the workflow to understand how user prompts are processed, mapped to support areas, and executed using available tools.

Explain how the Mem0 framework and Neo4J graph database are used to create a personal coach agent.

Outline the steps required to install and configure the agent, including setting up Docker, Telegram bot, and Engrok.

Identify the key steps in the video testimonial agent workflow, including form submission, persona selection, and video generation.

Analyze the process of concatenating video clips using FFmpeg, including installation methods and command execution.

Explain the AI agent sub workflows as tools in the video ad generation agent, including storyboarding, multi image generation, and multi video generation.

Revisit the process of concatenating video clips using FFmpeg.

Illustrate DSPy's declarative approach to defining system inputs and outputs.

Emphasize DSPy's self-improving capabilities through powerful optimizers.

Explains DSPy's core components, including language models, signatures, and modules.

Details evaluation metrics and optimizers, highlighting their role in improving performance.

Describe DSPy's quick start, including installing libraries and running a model.

Illustrate how DSPy optimizes prompts and exports them for use in other programs.

Identifies DSPy optimizers as a powerful tool for prompt engineering.

Evaluates the process of optimizing prompts to avoid jargon and improve performance.

Trains an AI model to tell jokes on specific topics.

Evaluates the AI's jokes using AI personas as judges.

Demonstrates how to teach LLMs to write articles using tool calls and DSPy.

Details the process of creating an agent that can create outlines and stitch together blog drafts.

Automates competitive monitoring by scraping and summarizing competitor websites.

Identifies new content and delivers executive summaries via email.

Constructs a financial analysis agent using Yahoo Finance and DSPy.

Demonstrates how the agent autonomously retrieves stock data and calculates returns.

Identify key invoice fields using a JSON format.

Analyze and normalize extracted data, improving accuracy.

Identifies relevant documentation for customer support inquiries.

Classifies queries and retrieves order information for refunds.

Illustrates building a research agent with web search capabilities using OpenAI.

Explains integrating the React module for tool calls and incorporating search results.

Constructs a life coach agent using memory and vector databases. Explains the setup, configuration, and usage of the agent, including memory search and addition.

Describe using Fal AI to embody agents with physical forms.

Illustrate a workflow for generating video responses with AI personas.

Demonstrates a video generation agent using AI.

Evaluates the agent's ability to create and score videos.

Learn the core concepts of LangGraph, a stateful orchestration framework for building AI agents. You'll explore nodes, edges, conditional flows, and shared state, while mastering patterns for routing as code, debugging, and visualization. This lesson teaches you the foundational skills needed to design robust, agentic workflows with human-in-the-loop capabilities.

Explore the building blocks of LangGraph, including nodes, graph state, and edges. You'll learn how to design atomic node functions, manage state with strong typing and reducers, and control execution flow using deterministic and conditional edges. This lesson teaches you the core mechanics behind orchestrating reliable, flexible agent workflows from start to finish.

Learn how to build and run your first LangGraph application by combining nodes, edges, and shared state. You'll set up typed state with reducers, define start and end nodes, and implement conditional edges to create dynamic workflows. This lesson teaches you the foundational quickstart patterns for constructing, executing, and debugging functional AI graphs.

Learn how to extend LangGraph agents with tool calling, conditional edges, and custom tool nodes. You'll implement tool integration with external APIs, design conditional routing functions for dynamic control flow, and manage advanced state updates with reducers. This lesson teaches you how to orchestrate looping workflows and build custom tool logic that powers real-world, agentic applications.

Learn how to build a document processing agent that extracts structured data from invoices using LangGraph and Pydantic. This agent demonstrates essential patterns for handling file uploads, parsing PDFs, and generating type-safe outputs. You'll master practical techniques for transforming unstructured documents into structured business data using AI.

Build an AI comedy agent that generates and evaluates jokes using Pydantic structured outputs and dual model support. You'll learn how to implement professional evaluation systems with multiple AI personas and create agents that iterate on creative outputs. This project teaches you advanced structured output patterns and quality assessment techniques for creative AI applications.

Create a sophisticated coaching agent with persistent memory using Mem0's hybrid vector and graph database. Learn how to implement long-term memory management, personalized responses based on conversation history, and apply proven coaching frameworks like Circle of Control and Helicopter View. This agent demonstrates memory-augmented generation patterns essential for building AI assistants that provide context-aware, personalized experiences across sessions.

Develop a content creation agent that writes professional marketing blogs using vector search and semantic retrieval. You'll learn how to implement RAG (Retrieval-Augmented Generation) patterns by searching a marketing corpus, managing blog outlines, and assembling final content with proper research grounding. This project teaches you how to build agents that create factually accurate, well-structured content by combining AI generation with knowledge base retrieval.

Build a professional customer service agent that demonstrates agentic file reading and dynamic documentation search. Learn how to implement smart query filtering, process refunds with business rules validation, and search orders across multiple parameters. This agent teaches you essential patterns for building production-ready customer service systems that maintain conversation context and handle real-world business logic.

Create an AI research assistant that transforms questions into comprehensive reports using a 5-stage DAG pipeline with concurrent query execution. You'll learn how to implement smart question clarification, focused research planning, parallel AI-powered search, and structured report generation. This project demonstrates advanced workflow orchestration patterns essential for building agents that perform thorough, multi-step research tasks.

Develop an automated competitor intelligence system that monitors competitor blogs, generates insights, and creates executive briefings. You'll learn how to implement web crawling, URL diffing with PostgreSQL, parallel content processing with Sparse Priming Representations (SPRs), and automated digest generation. This project teaches you production-ready patterns for building scheduled monitoring systems that track competitive landscapes and deliver actionable intelligence.

Build a multi-modal agent that plans, illustrates, and produces product advertisement videos by coordinating Gemini 2.5 Flash and fal.ai Veo 3. Learn how to generate photorealistic stills, create native video content, manage media-aware chat history, and automatically stitch scenes with ffmpeg. This agent demonstrates sophisticated media processing workflows and teaches you how to orchestrate multiple AI models for creative video production.

Create an end-to-end testimonial video production agent that generates persona-driven buying scenarios and produces professional video content. Learn how to implement structured prompt generation loops, coordinate with FAL's Veo3 API for video generation, and assemble final videos with ffmpeg. This agent demonstrates complex DAG workflows and teaches you how to build production video generation pipelines that create authentic, persona-based marketing content.

Explore documentation for N8N, DSPy, and LangGraph to engage with AI agent communities.

Design AI agents to maintain relevance in the evolving landscape of agentic AI skills.