Exploring XasCode & Generative AI – Event Recap

Three ETAS colleagues standing in front of a branded screen preparing for their presentation at the event

At 42Heilbronn, a cutting-edge coding school that follows an innovative peer-learning approach, we had the exciting opportunity to connect with students and discuss two of the most transformative topics in modern software development: XasCode and Generative AI for embedded systems. 42Heilbronn focuses on practical, project-based learning, preparing students for real-world challenges in the tech industry.

This event was not just about sharing knowledge but also about sparking conversations on the future of software engineering and safety-critical systems.

As part of our presentation, we demonstrated how Sphinx-needs and XasCode streamline requirements engineering and enhance traceability in complex development environments. Additionally, we explored how Generative AI, beyond chatbot applications, can be leveraged to support software engineering workflows, improve documentation, and assist in automated code generation.

Highlights

Demonstrating XasCode in action

We showcased how code-driven methodologies streamline software development, improve efficiency, and reduce complexity.

Exploring Generative AI for embedded systems

Our live demo illustrated how Retrieval-Augmented Generation (RAG) and fine-tuned AI models can support software engineering workflows.

Interactive discussions & future opportunities

Lively exchanges with participants, hands-on demonstrations, and discussions about potential collaborations and internships.

Sphinx-needs & XasCode

Managing complexity in software development

Safety-related software development demands precision, compliance, and structured workflows. During our session, we demonstrated how Sphinx-needs supports structured requirements management, ensuring that software projects adhere to strict safety and quality standards.
We also introduced XasCode, showcasing how software requirements, infrastructure, and configurations can be automated using code-driven methodologies. This approach allows software teams to enhance efficiency, reduce manual errors, and maintain a transparent version history throughout the development cycle.

"We bring cutting-edge technologies like GenAI and Docs-as-Code to life for the benefit of our customers. On the educational side, 42Heilbronn is an ideal partner, equipping students with the practical skills needed to turn these innovations into real-world impact."

Hendrik Post - Chief SW Architect & People Lead: Dependability & Data Solutions

Generative AI for embedded software

Beyond chatbots

Artificial intelligence is transforming software engineering, but its true potential goes far beyond chat-based applications. We explored how Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) can optimize embedded software development, providing solutions such as:

Intelligent documentation generation: Automating the creation of safety reports and technical documentation.
Automated requirement-to-code translation: Assisting developers in generating compliant and structured code.
Enhanced debugging and troubleshooting: Leveraging AI to analyze log files and suggest fixes in real-time.

Our live demonstration of a RAG-based AI assistant sparked a lively discussion on the possibilities and limitations of AI in regulated software development environments.

ETAS colleagues presenting to an engaged audience at an event.

Engaging with the next generation of software engineers

One of the highlights of the event was the interactive discussion with students, who brought insightful questions and perspectives on the practical applications of AI in software engineering. The interest in automation, AI-driven coding assistance, and best practices for handling complex software requirements was truly inspiring.