Quality Engineering meetup #14

In collaboration with Deutsche Bank, Qase is presenting the fourteenth Meetup on Quality Engineering in Berlin.

We’re bringing you three deep-dive talks: all designed to challenge the way you think about quality, testing, and collaboration.

As always, expect great conversations, snacks, and drinks. See you there!

AGENDA

1. Lilya Valeiko @ Deutsche Bank:
Testing bank products in the AI Era: From Survival to Quality Assurance

Working in a bank requires extra caution: the risks are real, and the consequences of failure can be serious. When AI helps your own developers produce ten times more code, QA has more to understand and verify. Now add a growing ecosystem of hundreds of tenant applications depending on the platform: every change can ripple outward, and the platform team is responsible for making sure nothing breaks downstream — even as those teams build faster and produce more code with AI too. Meanwhile, managers expect even more delivery. For QA, the mountain of things to understand and verify keeps growing.
What happens when the volume of change outpaces our ability to check it?

In this talk, Lilia will share how her team at Deutsche Bank is approaching this challenge. Their ambition goes beyond getting through the next batch of testing tasks: they want QA to help make the system better. Through the story of their approach, she'll explore what this growing complexity means for verification, how it challenges the way QA works, and what it takes to keep improving quality under increasing delivery pressure.

2. Aleksey Ogarkov @ Deutsche Bank: Ask the Codebase: Knowledge Graphs for Developers and Testers

In large organisations, understanding what other teams are building and how their code affects yours has always been difficult. With AI accelerating code production, keeping up is becoming even harder—for developers and testers alike. This talk introduces a knowledge graph built from 300 repositories, continuously updated from the code and searchable through natural language questions. We’ll look at how it helps developers understand dependencies and testers work out what to test, alongside plans to connect it to a test management system.

3. Vitaly Sharovatov @ Qase: 10× developers, 0.5× teams

A 10× developer sounds like a shortcut to a faster team. DORA’s 2025 research reports higher delivery throughput alongside greater instability as AI adoption increases. Faster code production can leave the team with more changes to check, more failures to investigate and more work to repair. An individual productivity gain can become a problem for the whole delivery system.

When faster engineering becomes a demand for more features, product and QA have less time to do their work properly. Product decisions arrive before the underlying problems have been thought through; QA has more changes to check against less clear expectations. Mistakes generate rework, leaving even less time to decide what is worth building and whether it works as intended.

The cost reaches development too. When they delegate coding, the agent does the work through which they used to learn the software. They can lose their grasp of the system and the reasons behind its design. With less practice, their ability to diagnose and change it can deteriorate. Working alone with agents also means less practice sharing knowledge, challenging assumptions and settling disagreements together. The risk is a team with more to check, less understanding to check it with, and fewer conversations to expose mistakes.
I’ll draw on my experience of mobbing with AI to explore how it could strengthen the whole team. Its coding speed could let product, engineering and QA stay involved together, guiding one agent through small changes. The aim is to preserve understanding and practise making decisions together while the software is being built.
A 10× individual gain is a failure if it leaves the team with more rework and a worse product. AI should help the whole team deliver something worthwhile and retain the skills to improve it.