Build AI Agents with LangGraph: Workflows, Memory & Streaming

Wishlist Share

About Course

In Build Your First AI Agent: Tools, Loops & Self-Correction, you learned how to build AI agents from scratch using Python. Now it’s time to take the next step: building structured, reliable AI workflows with LangGraph.

In this second course of the Zero-to-Hero Agent Roadmap, you’ll learn how to build graph-based AI agents, implement conditional routing, execute tasks in parallel, validate data using Pydantic, and improve performance with asynchronous Python.

You’ll also explore conversation memory using SQLite and PostgreSQL, and implement real-time streaming to create more interactive AI applications.

Through hands-on Python examples, you’ll develop the skills needed to build stateful, maintainable AI agents.

Capstone Project: Auto-HR Onboarding Agent

Put your skills into practice by building an Auto-HR Onboarding Agent that collects employee information, validates details using Pydantic, identifies missing data, and handles simulated HR approvals.

You’ll combine LangGraph workflows, conditional routing, persistent conversation memory, and real-time streaming into a complete application.

By the end, you’ll have a practical portfolio project demonstrating how to build structured, stateful AI agents.

Prerequisite: Basic Python knowledge. Completing Build Your First AI Agent: Tools, Loops & Self-Correction is recommended.

Show More

What Will You Learn?

  • Understand why pure Python agents face challenges in production and compare popular frameworks like CrewAI, AutoGen, and LangGraph.
  • Learn LangGraph fundamentals, including state, nodes, edges, and graph-based workflows.
  • Build intelligent tool-using agents with conditional routing, dynamic decision-making, and parallel task execution.
  • Validate user inputs and generate structured LLM outputs using Pydantic, custom validators, and output schemas.
  • Improve agent performance with asynchronous Python, async, await, and concurrent execution.
  • Implement conversation memory and persistent state using LangGraph checkpointers, MemorySaver, SQLite, and PostgreSQL.
  • Stream agent responses in real time, including workflow progress updates and live LLM-generated text.
  • Combine these concepts into a complete AI agent with structured data extraction, validation, conditional workflows, memory, and streaming.

Course Content

Module 1: Introduction to Agentic Frameworks & LangGraph
Understand why traditional Python agents struggle in production, compare popular agentic AI frameworks, and explore the fundamental concepts of LangGraph through practical examples.

  • Why Pure Python Agents Fail in Production
    05:17
  • Why Do We Need Agentic Frameworks?
    02:13
  • CrewAI vs AutoGen: Strengths and Limitations
    02:39
  • LangGraph: A Framework for Production-Ready AI Agents
    02:44
  • Understanding State, Nodes, and Edges
    03:07
  • Building Your First LangGraph Workflow
    04:52
  • Connecting, Compiling, and Running LangGraph Workflows
    05:23

Module 2: Advanced LangGraph Workflows — Conditional Routing & Parallel Execution
Learn how to build intelligent and efficient AI workflows using LangGraph. Explore conditional routing, LLM-powered tool execution, conversation state management, and parallel processing to create more flexible and scalable AI agents.

Module 3: Data Validation, Structured Outputs & Async AI Agents
Learn how to build reliable and efficient AI agents using Pydantic validation, structured LLM outputs, and asynchronous Python. Validate inputs, enforce output schemas, and improve agent performance through non-blocking execution.

Module 4: Memory & Persistence in LangGraph
Learn how to implement conversation memory and persistent state management in LangGraph. Explore in-memory checkpointing, SQLite database storage, and PostgreSQL persistence to build reliable AI agents that remember conversations across sessions and application restarts.

Module 5: Real-Time Streaming & Capstone AI Agent Project
Learn how to stream AI agent responses in real time using LangGraph and bring together everything you've learned by building a complete Auto-HR Onboarding Agent. Integrate structured data extraction, Pydantic validation, conditional routing, persistent memory, simulated approval workflows, and live response streaming.

Student Ratings & Reviews

No Review Yet
No Review Yet