Talks

Test automation in natural language, without code. Visual Language Models in conjunction with AI agents are already capable of this.

What are the challenges of testing non-deterministic systems based on LLMs? How is AI quality evaluated? What should be tested, and which frameworks can be used for this?

What should you do when a test fails? This talk will showcase how to use Playwright's tools to analyze errors, improve reporting, and create clear, actionable tests that help pinpoint issues more efficiently.

We will discuss not only the testing of desktop applications but also modern approaches to using neural networks in automation. We will evaluate how these technologies are changing the game by enabling the creation of automated tests without the need for programming skills.

We will discuss what design systems are and how they relate to interfaces, as well as how interface testing differs from regular testing. We will tell you how we tested without a tester and what we got out of it. And how we rethought our approach to testing with a QA joining us.

How you can use the detekt static analyser to improve the quality of code of backend autotests written in Kotlin.

I will briefly run through the areas, give some tips and useful links. We will practice answering questions and analyze some of them.

Generative model testing: what and why you need to know about the model and how to compose prompts for model hacking.