"AI Agents: Development Without Frameworks" is a practical guide to building robust and manageable AI agents using pure Python and direct interaction with AI APIs.
Course Advantages
Unlike most training programs, in this course, you won't rely on complex and opaque frameworks. You will learn an approach used by strong engineers and leading AI companies: minimal abstractions, full control over logic, and predictable system behavior.
Practical Application
During the course, you will analyze a real AI application that:
- processes text data (e.g., meeting transcripts);
- makes decisions on actions;
- calls external tools (creating calendar events, reports, etc.);
- forms a final result for the user.
In-depth Understanding of AI Agents
You will understand how AI agents truly work:
- what an agentic loop is and how it is implemented in practice;
- how language models are transformed into structured commands (JSON);
- how Python manages all logic, validation, and action execution;
- how to correctly design the system of tools;
- how to build reliable and secure AI applications.
Key Ideas of the Course
Special attention is given to the key idea of the course: the AI model does not execute code — it only suggests actions. All responsibility for execution, control, and security lies with your code.
Skills You Will Acquire
- create AI agents from scratch without LangChain and similar frameworks;
- work directly with model APIs (e.g., GPT or Claude);
- implement tool calls and process their results;
- build scalable and understandable architectures;
- adapt agents for any tasks—from automation to research.
Who Is This Course For?
The course is suitable for developers who want to move from "toy" AI applications to real engineering solutions and gain full control over their systems.
