Software Architecture: The Bad Parts

In this talk, I will demonstrate how so-called good practices combined with a shallow understanding of the domain can create a dangerous illusion of control.

Using a real-world example from the Connected Health domain—specifically, liver cancer risk alerting—I will show how a noun-driven design approach leads to excessive coupling and brittle systems.

The talk will explore common architectural pitfalls such as context violations, database coupling, domain leakage, and mixing read and write models. I will also challenge a popular industry belief by explaining why Clean/Hexagonal Architecture is not an architecture.

Attendees will leave with a clearer understanding of how design decisions shape system behavior—and how easily “best practices” can fail when the domain is misunderstood.

Mechanics of talking Must-have soft skills of Solutions Architect.

You hear it everywhere: if you want to be a modern Solutions Architect and not get left behind, you must know GenAI, DDD, EventSourcing, EventStorming, or EventModelling.
It’s only half the truth.
The other half—without which you simply won’t succeed—are soft skills.

I’ll share my non-technical experience working as a Solutions Architect and talk about:

- How to structure your statements
- What techniques to use to increase self-confidence during important meetings
- Why small talk matters
- How to impress a client
- How to present solutions clearly
- How to reduce stress levels

Soft skills are just like technical skills: you can improve them.

It’s not easy. Especially if you want to be more than an average Senior Engineer and keep your position safe in the coming years.

Events are domain atoms. Modernize your legacy system with event sourcing, FigJam, GenAI and fi

Ready to transform your domain modeling?

Event sourcing shifts focus from stale states to immutable facts—the true atoms of your domain. Instead of modeling systems around changing state, we model what actually happened: business events that remain valid even as rules and understanding evolve.

Many legacy systems become hard to change because their domain models drift away from business reality. By focusing on commands and events rather than nouns and data structures, we create thinner aggregates and architectures aligned with real business language.

This talk combines event sourcing, the decider pattern, and state-machine thinking with a practical modern workflow.
Domain discovery starts collaboratively in FigJam, evolves into executable specifications, and is continuously verified against the live codebase using MCP-powered fitness functions. GenAI supports exploration and refinement, helping teams keep architecture and implementation aligned.

Attendees will learn:

👉 Why commands and events improve reliability and adaptability
👉 How noun-driven models degrade architectures over time
👉 How the decider pattern simplifies domain logic
👉 How FigJam, MCP, GenAI, and fitness functions create continuous architectural feedback

A practical approach to modernizing legacy systems where domain models, architecture, and code evolve together — without rewriting history.

~ 50min