inflearn logo

Introduction to Building AI Agents Without a Framework

Build an AI agent directly with Python and a single while loop, without a framework. Add tool design, failures and costs, memory, human approval, evaluation, and a web interface one at a time, and ultimately complete a small coding agent that fixes its own failing tests.

1 learners are taking this course

Level Basic

Course period Unlimited

Python
Python
FastAPI
FastAPI
LLM
LLM
openAI API
openAI API
AI Agent
AI Agent
Python
Python
FastAPI
FastAPI
LLM
LLM
openAI API
openAI API
AI Agent
AI Agent

What you will gain after the course

  • Implement an agent loop that calls tools using only the OpenAI SDK and a while loop, without a framework.

  • Design tools that models use well: a tool registry, descriptions, and returning results concisely and structurally

  • How to prevent infinite loops, invalid JSON, and API errors with safeguards, and how to measure actual charges using usage.cost

  • Manage short-term and long-term memory using conversation history reduction, summarization, session storage, and note-taking tools.

  • A flow that shows the diff and obtains human approval before writing files or executing commands

  • Compare the models (DeepSeek and GLM) by measuring pass rate, average cost, and number of calls using an evaluation set.

  • A local web interface that displays real-time progress by episode using FastAPI and SSE

  • Complete a mini coding agent that reads failing tests, fixes them, and runs them again.

Recommended for
these people

Who is this course right for?

  • A developer who knows the basics of Python but has never built anything using an AI API.

  • People who have tried frameworks like LangChain and are curious about what actually happens inside them

  • For those who use coding agents like Claude Code and Codex and wonder, "How does this work?&quot

  • Those who want to build a small in-house automation agent themselves and control costs and failures on their own

Need to know before starting?

  • Python Basics: Functions, dict/list, for/while, try/except, f-string (Part 0: We’ll review them again with refresher.py)

  • Using `cd`, Running Commands, and Handling Environment Variables (`.env`) in the Terminal

  • It’s enough to have simply heard of HTTP APIs or JSON. It’s okay if you’re new to AI APIs.

Hello
This is dante

Career Verified

122

Learners

17

Reviews

4.5

Rating

2

Courses

Hello! I'm Dante.
I am the author of The Complete Guide to Vibe Coding Codex and Deep Dive into Inside React.

I have experience working as a software engineer at leading major corporations such as Coupang and Naver, and I have met with over 400 mentees to provide consulting on the practical application and learning methods of AI agents, as well as front-end development and learning strategies.

XL

The Complete Guide to Modern Vibe Coding Codex

If you are interested in my mentoring,

Please check https://mentoring.inflearn.com/mentors/1535 :) nhé :)

스크린샷 2025-08-14 오후 9.49.00.png.webp

Inside React: Deep Dive Again

Kyobo Book Centre https://bit.ly/4mbixaE

Aladin https://bit.ly/4m9tOYO

Yes24 https://www.yes24.com/product/goods/151041836

More

Reviews

Not enough reviews.
Please write a valuable review that helps everyone!

dante's other courses

Check out other courses by the instructor!

Similar courses

Explore other courses in the same field!

Limited time deal

$912,952.00

50%

$68.20