AI Coding Tools - The developerʼs perspective

In the second volume of our AI-focused series, we sat down with two Netminds specialists to hear about their hands-on experiences with AI coding assistants. Meet Vitalii, a Senior Java/Python Engineer, and Roman, a Senior AQA Engineer - teammates who not only use AI in their daily coding and workflows but also integrate AI and LLM features directly into the software products they build.

"I use ChatGPT and Augment Code in my daily work, and they’ve become as essential to me as my IDE.

Augment Code’s greatest strength is its deep understanding of project context, which allows it to handle repetitive but widely distributed changes in the codebase with remarkable efficiency. This capability not only saves time but also reduces the risk of missing small but important updates scattered across different modules. Another area where Augment Code shines is in simplifying the process of writing tests, making test coverage a far less tedious task. I also rely on AI as a brainstorming partner — it’s great for surfacing alternative solutions or approaches I might not have considered.

Of course, I’ve learned that AI’s output always needs a second look because of “hallucinations” — situations where it produces convincing but factually wrong or irrelevant responses.

In my current project, we’re also building our own MCP servers in Java and Python that integrate directly with our services and can be accessed via LLMs without a dedicated UI. This opens up exciting new ways for teams to interact with internal systems, but it comes with its own challenges. One key drawback is the need to craft a robust pre-prompt, as the quality of responses depends heavily on its clarity.

Overall, I see huge potential for AI tools in software development and believe they will become a standard part of every developer’s toolkit in the near future."

Vitalii Fedyna, Senior Java/Python Engineer at Netminds 

"As an Automation Engineer with 8 years of experience, I have gained extensive hands-on expertise with a variety of AI-powered tools that assist in both software development and test automation. My professional toolkit includes ChatGPT, Windows Copilot, Copilot for IntelliJ IDEA, and AI-driven automation platforms such as UiPath with AI, ChatUniTest, and ACCELQ Autopilot.

From my experience, Windows Copilot is a convenient productivity tool for everyday tasks within the Windows ecosystem. It helps streamline workflows and provides quick assistance, but when it comes to advanced problem-solving or context-aware development, it still lags behind the capabilities of ChatGPT and newer large language models.

In contrast, ChatGPT stands out as a more powerful tool for reasoning, contextual understanding, and generating high-quality solutions beyond simple task execution. Similarly, Copilot for IntelliJ IDEA is useful in coding scenarios—especially when switching between different testing projects—since it helps recall context and speeds up implementation.

I have also explored AI-driven test automation solutions such as ChatUniTest and ACCELQ Autopilot, which aim to automatically generate and self-heal test cases. While these tools are promising, I believe they are still in an early stage of maturity. At this point, the generated code and the overall usability of these platforms often feel inconvenient and inefficient compared to more established approaches.

In summary, my experience has shown me that while Windows Copilot is helpful for general productivity, and Copilot for IntelliJ IDEA assists with code-level tasks, ChatGPT currently delivers the most reliable and advanced support for complex problem-solving. On the other hand, AI-driven test case generation tools are still evolving and not yet ready to fully replace traditional, well-structured test automation practices."

Roman Fedun, Senior AQA Engineer at Netminds 

AI has already secured its place in the daily tasks of developers and testers. But the real question is: what’s next? Will it become a true “teammate” you can’t work without, or remain just a convenient assistant?

What do you think - what role will AI play in your work a few years from now?