SciEngineer

04/06/2026

Most SDV programs don't fail in testing. They fail in timing.

In many automotive development programs, defects are introduced during requirements and design but only discovered much later, when fixing them is significantly more expensive.

In his Software Testing Conference 2026 presentation, our Application Engineer colleague Bence Nagy explored how shift-left verification, Model-Based Design, simulation, virtualization, and AI-assisted engineering workflows can help close this gap.

The full material is now available as a white paper:

Shift-Left in the V-Cycle: Digitalization and Virtualization in Software Verification

Download here:
https://sciengineer.com/shift-left-in-the-v-cycle-digitalization-virtualization-in-software-verification/

A practical shift-left approach using Model-Based Design and simulation to reduce late-stage defects and improve verification efficiency.

27/05/2026

A great example of AI solving practical grid and energy management problems, not just demos.

Read the full story: https://www.mathworks.com/company/user_stories/vgen-develops-virtual-power-plant-with-deep-learning-and-machine-learning.html

20/05/2026

Tomorrow at the Software Testing Conference 2026, our colleague Bence Nagy will be speaking about one of the most important challenges in automotive software development: how to detect defects earlier in the lifecycle.

His talk, “Shift-Left in the V-cycle: Digitalization and Virtualization in Software Verification”, focuses on how software-defined vehicles are changing the way we design, develop, and verify complex systems.

In modern automotive projects, the real challenge is not only the number of defects, but when they are discovered. Many issues still appear late in integration or testing phases, where fixing them is significantly more costly.

Bence will explore how Model-Based Design, simulation, virtualization, and AI assisted engineering workflows can help bring verification earlier, closer to where issues are introduced, improving both efficiency and predictability.

More about the event:
https://businessevents.hu/rendezvenyek/konferencia/szoftverteszteles-konferencia-2026

Tesztelési stratégiák az AI-vezérelt automatizáció és a folyamatos szállítás metszéspontjában Bevezető Program Előadók Kinek ajánljuk? Szponzoráció és kiállítás PDF letöltés Bevezető A...

14/05/2026

As Generative AI continues to reshape engineering workflows, integrating AI into Model-Based Design is becoming increasingly important.

In this upcoming webinar by MathWorks, discover how Simulink® Copilot and the Simulink Agentic Toolkit help engineers accelerate development, automate workflows, and enhance productivity using Generative AI within Simulink®.

You’ll learn how AI-powered assistance can support model understanding, troubleshooting, task automation, and intelligent engineering workflows through the Model Context Protocol (MCP).

If you're working with simulation, control systems, embedded software, or Model-Based Design, this session will provide practical insights you can apply directly to your engineering processes.

Register for the event here:

Generative AI accelerates the design, simulation, and testing of engineered systems. It can be applied in different ways, including interactive assistance in your development environment and integration with external AI agents. When combined with trusted tools such as Simulink and Model Based Design...

12/05/2026

An inspiring few days at CASSINI Hackathon Budapest - filled with creativity, collaboration, and ideas with real-world potential.

Great to see such a strong community of innovators shaping the future together. We at SciEngineer are proud to have been one of the sponsors supporting this inspiring event. 🚀

07/05/2026

Agentic AI is starting to change how we think about model-based engineering workflows.

Our engineering colleague, Bence Nagy, recently tested the new Simulink Agentic Toolkit on a real engineering task: DC motor modelling and control verification.

The goal was to build a validated Simscape model of a DC motor with H-bridge control for speed and direction testing.

In preparation, he also followed the official MathWorks guide on Agentic AI with MATLAB, see it in the first comment.

The result was surprising in its simplicity and speed:

In less than 20 minutes, a working Simscape model was generated that:
• matched datasheet behaviour (~2200 RPM no-load speed)
• supported CW and CCW direction control
• responded correctly to PWM inputs (~1100 RPM at 50% duty cycle)
• was ready for control algorithm testing

What stood out most was not only the speed of generation, but that the first model already worked without extensive debugging.

This is an interesting early example of how AI-assisted engineering workflows may support faster prototyping and validation in model-based engineering and automotive software development.

28/04/2026

Great to be back at the NAFEMS Conference, especially this year in Budapest, our home base. It’s always a highlight for us to be part of this community, and having it here, in Hungary makes it even more special.

Over the next two days, our team is at the stand showcasing simulation-driven solutions with COMSOL Multiphysics and sharing how we approach complex engineering challenges.

Our expert, Arpad Forberger, PhD, is also on site to share insights and discuss your challenges.

If you’re attending, come by, take a look, and let’s connect. 👋

24/04/2026

🚀 The Cassini Hackathon has officially kicked off!

We’re proud to be sponsoring this event, bringing together innovators, engineers, and researchers to tackle real-world challenges using space data.

There’s huge potential for innovation in this space. From protecting our water resources to building the next generation of environmental start-ups, the opportunities are just beginning.

A big thank you to the organizers for screating such a meaningful and forward-looking platform.

Best of luck to all participants—we’re excited to see your ideas take shape! 🌍

23/04/2026

34% faster compilation? Real results from the field.

In automotive software development many teams are excited about AI driven workflows but some of the biggest efficiency gains still come from small practical improvements available today.

In one of our recent projects our application engineer Bence Nagy applied a new built in Simulink feature the Model Componentization check in R2026a to a large automotive model.
The result? A 34% reduction in model compilation time with no extra infrastructure needed.

This kind of optimization can save engineering teams hours of waiting time every week and lets them focus on iteration instead of builds.
If you work with large Simulink models in automotive or SDV related projects this is definitely worth exploring in your own environment.

More details about Simulink Performance Advisor Checks
https://www.mathworks.com/help/simulink/slref/simulink-performance-advisor-checks.html

Describes Performance Advisor checks for Simulink software.

22/04/2026

The 2026 DATA-ENG Hackathon wrapped a few days ago, and there’s a lot worth looking back on.

In our previous post, we shared a few glimpses of the contest—teams deep in discussion, working through the modeling challenge under time pressure. The task, brought by AUMOVIO, focused on modeling the speed difference between the front and rear wheels of a motorcycle using hashtag and .

Now we’re back with photos from the finals and presentations. Same problem, different solutions, and each team approached it in their own way. From an industry side, this kind of fresh thinking and variety is always valuable.

For us at SciEngineer, this is something we care about: bringing academic knowledge and real industry problems into the same space, so both sides can learn from it and get value out of it.

We would like to thank all teams for their work and congratulate the winners.

Thanks also to AUMOVIO for their long-term partnership and co-organizing the event with us, and to MathWorks for sponsoring it. Special thanks to University of Pannonia for hosting the hackathon and providing MATLAB Campus-Wide Access, which ensured students came prepared and comfortable with MATLAB and Simulink.

Learn more about Campus-Wide Access here: https://sciengineer.com/academic-use/matlab-campus-wide-license/