iSAQB® Architecture in the Context of Agent-Based Software Development (AGENTA)
AI Guidelines, Context Engineering, and AI Governance for Architects (CPSA Advanced Level)
.svg)
Machine Learning Engineer
In the AGENTA training course, you'll learn how generative AI and agents support and transform architectural work. You'll explore AI-driven architectural decisions, the provision of architectural knowledge to agents, and appropriate control mechanisms for quality and architectural specifications.
In addition, you'll learn how AI can be used to analyze existing systems and what requirements this entails for governance, security, traceability, and accountability.
1. Introduction and basics
- Fundamentals of Generative AI, Large Language Models, Prompt Engineering, and Context Engineering
- How AI Agents Work and Their Applications in Software Development and Architecture
2. AI-Driven Architectural Decisions
- Analyzing Quality Goals, Architectural Drivers, and Structural Weaknesses Using AI
- Develop and evaluate architectural variants, and validate decisions through experimentation
3. Provide architectural knowledge to agents
- Preparing architectural decisions, quality goals, rules, and domain knowledge for agents
- Make architectural knowledge available in a targeted and effective manner using appropriate formats and mechanisms
4. Align agents with architectural goals
- Ensure that architectural decisions and technical specifications are supported by appropriate control mechanisms
- Make targeted use of architectural tests, static analyses, quality gates, and continuous verification
5. Extract architectural information from existing systems
- Extracting architectural information from source code, configurations, tests, and interfaces
- Create and keep architectural renderings and documentation up to date with AI support
6. Governance and Quality Gates in the Use of AI
- Take into account data protection, the legal framework, traceability, and organizational responsibility
- Reduce risks associated with prompt injection, tool usage, and sensitive information by implementing appropriate protective measures
7. Further Development of the Software Architecture
- Maintaining Responsibility and Architectural Judgment When Using AI and Coding Agents
- Shaping Changes in Roles, Collaboration, and Architectural Work in AI-Enabled Organizations
Requirements:
Have experience in software development or software architecture, as well as a basic understanding of typical architectural concepts and architectural artifacts.
Practical experience in software development or software architecture; a basic understanding of software architecture and typical architectural artifacts; basic knowledge of modern software development processes and collaboration within development teams; a basic understanding of source code repositories, build pipelines, and common development tools. Experience with architectural decisions, quality requirements, architectural documentation, as well as continuous integration and delivery, testing, and code reviews is also helpful. A basic understanding of generative AI and large language models, initial experience with AI assistants or coding agents, and an awareness of the typical limitations of AI systems will make it easier to get started. Experience in translating business or quality objectives into technical decisions, as well as an understanding of domain concepts, business rules, and stakeholder communication, is also advantageous.
Technical requirements:
There are no specific requirements for the training environment, as the exercises are conducted using cloud- and web-based tools such as Miro and Draw.io, as well as platforms like Jupyter and HuggingFace. All that is needed is a stable and sufficiently fast internet connection.
- You can specifically integrate AI and agents into your architectural work.
- You'll gain confidence in dealing with AI-driven architectural decisions and control mechanisms.
- You'll understand how architectural knowledge can be made useful for agents and how architectural work can be made transparent.
- You can make informed assessments of the opportunities, limitations, and responsibilities associated with the use of AI in software architecture.
Hands-On Learning: You'llexperience practical methods and techniques through realistic challenges in software development and architectural work.
Direct Application in Practice: You’llapply what you’ve learned to typical situations in your own projects, teams, and development processes.
Holistic perspective: Youdon't view the use of AI in isolation, but rather in conjunction with architecture, quality, governance, and accountability.
- Visual collaboration through, for example, whiteboarding
- High level of interactivity through customized exercises and breakout rooms
- Ideal trainer support even in the breakout rooms
Credit Points for CPSA-A Certification:
Through the AGENTA training program, the " attendees " earn 10 technical and 20 methodological credit points under the iSAQB Advanced Level Program.
.svg)