“You don’t appreciate the effectiveness of AI until it’s taken away” — interview with an author of the EngX AI-Supported Testing course

“I wanted a quiet and calm profession when I grew up”

Tell us about your childhood: where were you born and how did you choose your profession?

— I was born and live in Kyiv, Ukraine. I feel at home here, so I never wanted to move to another country.

In grades 10–11, I worked as a cloakroom attendant — I had a key, and I could give and take people’s outerwear. I liked this job because it was quiet and calm. That’s how I earned my first money.

As a child, I dreamed of becoming a pathologist: again, a profession without deadlines and hurry. But my parents decided that being a pathologist was unpromising, and the future was information technology. They took my documents to the Department of Information Technology at the National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute.” First, I got a master’s degree and taught in my department, and later, I obtained a PhD in Computer Science.

“I make the client happy”

What was your career path?

— In 2012, I started working part-time at EPAM as a data analytics trainer in the R&D area. Soon afterward, the Head of the Testing Lab position became available. I was offered that position, and I switched to full-time. I worked in R&D for 4 years and, in 2016, I joined testing production projects.

Now, I am a test manager — I set up testing processes. I establish and analyze metrics to assess product quality to make the client happy.

What inspires you in your work?

— I especially appreciate the opportunity to learn something new and apply it in practice. I love working with diagrams and I regularly improve my skills in business analysis. Before, it took a lot of time to draw all the elements of a model, and now you can experiment with AI tools and generate a script that will draw the model.

I look for courses on LinkedIn Learning and on the websites of major AI solution manufacturers — Microsoft, AWS.

“Don’t do work-related tasks in the maternity hospital”

Do you have any hobbies?

— I like to crochet, knit, and cross stitch. And relax in nature: to read a fantasy book, drink delicious tea, and watch the surrounding world.

Do you manage to maintain a balance between work and life?

— There are difficulties. I didn’t take maternity leave for my first two children, but I decided to try it a month before my third child was born. My phone began ringing persistently while I was giving birth. The doctor thought it was a worried husband. I answered the call and found it was a colleague with an “urgent” question. Don’t repeat my mistake: do not do work-related tasks in the maternity hospital.

“Writing prompts is a skill that can be developed”

How did you come up with the idea of creating the EngX AI-Supported Testing course?

— I thought about the course about 1.5 years ago, when the GPT-3 model first appeared. I started using it for test tasks — generating test cases and test data sets. ChatGPT coped well with the tasks, but colleagues did not share my enthusiasm.

When creating a request, it is important to incorporate all aspects that may affect the result. My colleague wrote a prompt “Create a Page Object for Wikipedia.” ChatGPT replied that it could not do so but gave an algorithm for creating one. This is a useless result for a person who knows how to create a Page Object but wants to make their work easier. I changed the original prompt and added context about the required programming language — “Create a Page Object for Wikipedia in Java.” After that, everything worked out.

I realized that writing prompts is a skill that can be developed. And I wanted to systematize my knowledge to pass it on to my colleagues.

Who helped create the course and what difficulties did you face?

— Oleksandr Antonov, Software Testing Team Leader, and Alexey Shcheglov, Project Manager at EPAM and EngX expert, helped me. I have been working on the same projects as Oleksandr for many years. We worked together to select and structure the course materials. Alexey reviewed our content, left professional comments, and suggested ideas.

The rapid development of GenAI added difficulties to the creation of the course. While we were working on the materials, the models and tools changed. Sometimes, we had to completely rebuild the content, so that the course remained relevant. But it was interesting to explore new tools.

“AI frees testers from routine tasks”

How does the EngX AI-Supported Testing course differ from other courses on working with AI?

— There are courses on techniques, approaches, and frameworks in prompt engineering with examples of generating a pie recipe and a children’s party script. Our course is focused on solving routine testing tasks with AI. Specialists can spend more time writing interesting, more complex tests.

Who is the course suitable for?

— The examples in the course are aimed at manual testers. For automated testers, it also includes materials on generating test data and writing scripts to create pipelines in Jenkins and deploy virtual machines in the cloud.

“Productivity indicators decreased by 30% without AI”

Is there any objective confirmation of the benefits of AI for solving work tasks?

— I was on a project that used Copilot. The whole team said that it didn’t really help with the work. Then, the client stopped paying for the license and the team’s productivity indicators dropped by 30%. AI can fit so organically into work processes that its effectiveness is not appreciated until it is taken away.

To get the benefit of using AI tools, you need to invest your time in learning them. Experimenting will help you form your own approach to working with them — and is the most productive way to incorporate them into your workflow.

What advice do you have for those who have not yet started using AI in their work?

— You can try to introduce it into your life: ask it to be a guide during a trip or to calculate the calories in your meal. This will help you see that AI really copes with the tasks. After that, it will be clearer how to use GenAI for work tasks.