Guide to Agentic AI – Build a Python Coding Agent with Gemini

The course is an excellent practical introduction to building AI agents, focusing on the agentic loop and tool calling. Lane Wagner’s teaching style is clear and methodical, making complex concepts accessible. The hands-on approach ensures that viewers not only understand the theory but also gain practical experience in implementing an agent from scratch. The use of the free Gemini Flash API is a significant advantage, lowering the barrier to entry. The course covers essential aspects such as project setup, API integration, and the implementation of four core tools, which are well-explained with code examples. The security warning about running AI-generated code is a responsible addition, highlighting potential risks. The content is technically accurate and aligns with current best practices in AI agent development. However, the course does not delve into advanced topics like multi-agent systems or fine-tuning, which might be expected by more experienced developers. The sources cited are primarily the boot.dev platform and the Gemini API documentation, which are reliable but limited. The adéquation between title and content is strong, as the title accurately describes the guide. Overall, the course is a valuable resource for developers looking to understand and build AI agents, with a high level of practical utility.

The title accurately reflects the content: a guide to building a Python coding agent using Gemini, with a focus on agentic AI concepts.

8/10

The tutorial is well-structured, hands-on, and based on a reputable platform (boot.dev) and API (Gemini Flash). It provides practical implementation details and includes a security warning about running AI-generated code. Sources are limited to the course itself and the boot.dev platform, but the content is technically accurate and reproducible.

• 00:00Introduction
• 01:14Why Build an AI Agent?
• 01:49Course Overview & What We're Building
• 02:25How to Follow Along
• 03:47What is an AI Agent? (Agentic Loops & Tool Calling)
• 06:03The Agent's Four Tools
• 07:58Prerequisites & Project Goals
• 10:08Demo: Agentic vs. One-Shot Responses
• 13:07Python Project Setup with UV
• 15:44Getting Started with the Gemini API
• 19:21Making Your First API Call
• 24:44Accepting Command-Line Arguments
• 27:46Managing Conversation History
• 30:39Adding a Verbose Flag for Debugging
• 33:35Setting Up the Project for Our Agent (Calculator App)
• 36:23Building Tool #1: Get Files Info
• 49:39Building Tool #2: Get File Content
• 58:24Building Tool #3: Write File
• 1:05:26Security Note: Dangers of Running AI-Generated Code
• 1:07:30Building Tool #4: Run Python File
• 1:18:00Understanding the System Prompt
• 1:33:10How Tool Calling Works: Declaring Functions for the LLM
• 1:41:38Adding All Function Declarations
• 1:49:19Implementing the Function Calling Logic
• 1:57:30Creating the Agentic Loop
• 2:07:11Final Demo: Agent Fixes a Bug Autonomously
• 2:13:44Conclusion & Next Steps

• freeCodeCamp News — General resource for articles and tutorials.
• Scrimba — Sponsor link, interactive learning platform.
• boot.dev Course: Build an AI Agent in Python — Interactive version of the course.
• freeCodeCamp — Main platform hosting the course.

• boot.dev Course: Build an AI Agent in Python — The interactive version of the course, providing additional practice and exercises.