Automate Dev with Self-Testing AI Agents

ABOUT THE COURSE:

The course "Automate Dev with Self-Testing AI Agents" is dedicated to creating a reliable development process using AI agents that not only make changes to the code but are also capable of checking the results of their work. While modern AI-based coding tools have high speed, they cannot see the actual behavior of the interface after changes — this course addresses exactly this gap.

Working with Playwright

You will learn to use Playwright as a verification and observability layer that allows you to check UI behavior, identify regressions, and form reliable artifacts from a running application. As part of the course, you will create a full-fledged verification harness for agents: deterministic tests with embedded acceptance criteria, a proof pipeline (traces, screenshots, logs, DOM snapshots), and a secure tool for agent interaction with the UI with clear boundaries and control.

Practical Focus

The course is practice-oriented: you will explore how to ensure test reproducibility, quickly diagnose failures using structured reports, and integrate all these processes into CI with clear artifacts. The approach is independent of a specific model or editor and is built around Node.js and TypeScript.

Course Outcomes

The result will be a ready verification infrastructure that makes UI changes verifiable, explainable, and reliable — both for engineers and AI agents.