Agentic Software Engineering - Michael van Engelshoven, Torben Keller - Deutsch
Generative AI is changing how we develop software - from individual code snippets to complete, agent-driven development processes. In this three-day training course, you will learn how to harness this transformation for your team.
You will learn how to use AI assistants and autonomous coding agents in a targeted manner: From intelligent requirements analysis and automated architecture design to autonomous development agents that implement complex features. You will always remain in control and understand where human expertise is indispensable.
After the training, you will be able to assess which AI tools and approaches are suitable for your specific challenges. You will know how to gradually move from simple code completion to advanced agent-based workflows.
Agenda
- introduction & context
- Why foundation models and generative AI are shaping the future of software development
- Technological change and its concrete impact on your day-to-day work
- Clarify terms: LLMs, multimodal models, foundation models
- Requirements with generative AI
- AI-supported creation, analysis and refinement of requirements
- Generating user stories from natural language and multimodal content
- Automatically translate customer needs into structured requirements
- AI-supported classification and prioritization of requirements
- Create prototypes faster and effectively involve non-technical team members
- design architecture efficiently
- Automated derivation of system architectures from requirements
- Create software design proposals using generative AI
- Reflecting and documenting architectural trade-offs
- Agent-based simulation and evaluation of architecture variants
- Accelerate implementation
- Increase efficiency from code completion to AI-driven implementation of complete features
- Support for API integration and library usage
- Increase code quality through automated refactoring recommendations and documentation
- Understand and modernize legacy code faster
- optimize testing & quality assurance
- Generate tests automatically: Unit, integration and end-to-end
- Create synthetic test data and identify edge cases
- AI-supported reviews for test coverage and quality
- Make CI/CD more efficient
- Generate CI/CD configuration automatically (GitHub actions, GitLab pipelines)
- Automated creation of release notes and changelogs
- AI-based security scans and compliance checks
- Self-healing pipelines and automated troubleshooting
- improve operation & monitoring
- Accelerate incident management with AI-based root cause analyses
- Automated monitoring and prioritization of alerts
- AI-based log analysis and integration with observability stacks
- Strengthen maintenance & further development
- Automated bug fixing and ticket management
- AI-based explanations of code changes (“Diff Explainers”)
- Automatic detection of regressions and technical debt reduction
Your Trainers
Torben Keller
INNOQ
Torben has several years of experience in the development, maintenance, and operation of applications. He is currently focusing primarily on the Spring Framework, moduliths, the Vue.js web framework, and data contracts.
Michael van Engelshoven
INNOQ
Agentische Softwareentwicklung, Frontendarchitektur
Michael van Engelshoven is a senior consultant at INNOQ. His work focuses on the development of modern web applications with a clear emphasis on front-end architectures and contemporary web technologies. He attaches great importance to well-structured code bases, clear architectures and a smooth developer experience – from the first commit to the running application. In addition to his work in front-end development, he is intensively involved in agent-based software development and the integration of AI-supported assistants into technical and organisational processes. He is particularly interested in how such systems can productively support teams and open up new possibilities for interaction and automation. Topics related to Git workflows and traceable commit histories are also close to his heart – because clean collaboration starts with version control.