ArchitectureConference50min
Software Design as Code - from Concepts to AI-driven Automation
The session demonstrates how to modernize software design by treating it as code—versioned, automated, and integrated with development workflows. It covers generating and maintaining diagrams and architecture artifacts via tools like Mermaid and Structurizr, and shows how AI can automate documentation, synchronize designs with code, and streamline architecture evolution.
talk.summaryAiDisclaimer
Marcin SamsonowskimBank
talkDetail.whenAndWhere
Thursday, June 18, 13:25-14:15
Room 4A
talks.roomOccupancytalks.noOccupancyInfo
Designing software no longer needs to be a manual, disconnected, and documentation-heavy. This session shows how to treat design as code - versioned, automated, and integrated into your development lifecycle.
We’ll move from simple markdown-based sketches to structured notations like UML, BPMN, and C4, all generated and maintained as code with tools such as Mermaid, PlantUML, Structurizr, etc. Learn how architecture decisions or diagrams can be tracked, validated, and evolved automatically.
Finally, we’ll explore how AI takes this even further - generating diagrams, ADRs, and documentation, and keeping designs in sync with your codebase. Real-world examples will illustrate how AI-assisted tools and LLM-based solutions (like ChatPRD) can support smarter documentation and architecture evolution.
Whether you’re a developer, architect, or DevOps engineer, you’ll leave this session with practical insights into:
- Choosing the right level of design automation for your context
- Integrating diagrams, ADRs, and docs directly into your codebase and CI/CD
- Leveraging AI to close the loop between design, implementation, and documentation
Join us to discover how to make your software design process as agile, automated, and intelligent as the code you write.
We’ll move from simple markdown-based sketches to structured notations like UML, BPMN, and C4, all generated and maintained as code with tools such as Mermaid, PlantUML, Structurizr, etc. Learn how architecture decisions or diagrams can be tracked, validated, and evolved automatically.
Finally, we’ll explore how AI takes this even further - generating diagrams, ADRs, and documentation, and keeping designs in sync with your codebase. Real-world examples will illustrate how AI-assisted tools and LLM-based solutions (like ChatPRD) can support smarter documentation and architecture evolution.
Whether you’re a developer, architect, or DevOps engineer, you’ll leave this session with practical insights into:
- Choosing the right level of design automation for your context
- Integrating diagrams, ADRs, and docs directly into your codebase and CI/CD
- Leveraging AI to close the loop between design, implementation, and documentation
Join us to discover how to make your software design process as agile, automated, and intelligent as the code you write.
Marcin Samsonowski
Architect and engineer with a passion for continuous learning and knowledge sharing. Focused on people, processes, and finding the right balance between quality and effectiveness.
Having worked across various industries, company sizes, and roles - from startups to large enterprises - he thrives on connecting a strategic perspective with practical execution. His main interests lie in Enterprise and Solution Architecture, where he applies a holistic view to make technology serve business goals effectively.
Outside of work, he’s passionate about music - singing, playing, and dancing - as well as sports ranging from martial arts to snowboarding.
Having worked across various industries, company sizes, and roles - from startups to large enterprises - he thrives on connecting a strategic perspective with practical execution. His main interests lie in Enterprise and Solution Architecture, where he applies a holistic view to make technology serve business goals effectively.
Outside of work, he’s passionate about music - singing, playing, and dancing - as well as sports ranging from martial arts to snowboarding.