AI agents develop quickly—but without you, they develop the wrong way. Learn how you, as a software architect, can stay in control in the age of coding agents: through targeted context engineering, architectural guardrails, and governance that really works.
AI agents and generative AI are fundamentally changing how software is developed. But they don’t eliminate the need for good architecture they make it even more urgent.
In this hands-on training, you’ll learn how to work effectively as a software architect in an agent-driven development environment: how to make architectural knowledge accessible to agents, how to design and enforce architectural guidelines, and how to maintain control over quality, governance, and decision-making despite a high degree of automation.
Agenda
- How do I extract architectural documentation from the code using AI? What can an agent not determine?
- How do I evaluate architectural solutions created using AI?
- What challenges arise when providing architectural information to agents?
- What architectural information does an agent need?
- What hard requirements does an agent need?
- How do I evaluate and prioritize guardrails?
- What security risks do agents pose?
- How does agentic engineering impact the environment and compliance?
- How is the discipline of architecture changing due to AI?
Audience
This training is designed for software developers and software architects who work in organizations that use generative AI and agents as tools in software development or plan to do so. It is particularly well-suited for anyone who wants to understand how to maintain architectural control in AI-powered development environments. Ideally, you have already used a coding agent before.
Training Objectives
How AI agents and LLMs work and what implications this has for architectural practice
How to use AI to make data-driven and experimental architectural decisions
How to prepare and provide architectural knowledge in a way that’s suitable for agents
How to keep agents within architectural boundaries using preventive and detective guardrails
How to extract and process architectural information from existing systems with AI support
How to ensure governance, legal compliance, security, and traceability when using AI
How roles, teams, and architectural work itself are changing in the wake of AI
Your Trainers
Niklas Trentmann
embarc
Software Modernization, Agent-Based Software Development, Software Architecture for Distributed Systems
- Software architecture for agent-based development
Niklas Trentmann advises and supports teams and companies in effectively integrating software architecture, AI, and modern development processes. With over 10 years of experience in backend development and distributed systems, he understands how crucial well-designed architecture is for stability and scalability. He has a particular interest in Agentic Development. Niklas combines technical expertise with organizational experience: He has built teams, led transformations, and created structures in which autonomous teams can work reliably and with a focus on quality.