Systematic

Intelligent software to empower first responders. Our software Rezilient Responder is built to ensure fast response in case of fire and rescure. Clear directions to the location of the incident. Continous updates on incidents and access to vital information at the scene to keep the first responders efficient and safe. Follow the link in comments to read more about Rezilient Responder.

Every second counts when a fire and rescue team is on the way to an incident. That is what Rezilient Responder is built for. Updates follow as the incident develops, so teams can have the latest information before they arrive. A map showing relevant information and the rescue sheet available directly from the vehicle. Watch the video to see it in action and find more information via the link in the comments.

Is keeping the human in the loop in every battlefield decision a requirement or a risk? In a combat situation, speed decides the outcome. If an adversary's AI allows for more autonomy and a shorter decision chain, waiting for human sign-off can cost us time we may not have. But where does this lead us – as a society and as an industry developing software used in life-and-death situations? I put that question to the room at Computerworld Danmark's Cloud & AI Festival 2026 yesterday afternoon. In my opinion, the answer is not to remove the human. There is always a person involved somewhere. The question I wanted to discuss was how close that person needs to be to each decision. And this cannot be answered without being mindful of data quality. The quality of data varies greatly depending on its source. Is it coming from the web and social media, or from confidential, verified intelligence services? The level of trustworthiness varies greatly. If we want the AI to have more responsibility, we need to know what data it is reasoning from. That is why we design our solutions to factor in these aspects, so the human who ultimately makes the decision knows what the AI is thinking and the quality of the data behind the AI agent's recommendation. I do not think there is a definitive answer to how much the human should be in the loop. Not today at least. But the sequence matters. What the system reasons from has to be settled before we decide how much room to give it.

Remember the APEX Team? They were feeling a bit down because they were missing a key member—a dedicated QA Engineer. https://lnkd.in/exRHnc_m But look at them now! 😊 As of August 17th, we welcomed Phillip Løvenhardt Vincent Nielsen to the team as our dedicated QA Engineer. Phillip is not only bringing joy back to the team but is also playing a crucial role in our development process. He ensures quality assurance from the very beginning, helping us deliver world-class software here at Systematic Welcome, Phillip! The difference you've made is already incredible. 🎉 #NeverStopDeveloping

Everyone wants more AI in Defence, but almost no one has enough data to train it. I lead a team building systems including agentic AI for the Defence domain: computer vision, data pipelines, and autonomous workflows that hold up in the most demanding real-world conditions. And the most demanding condition of all is not having enough data. My colleague @Julian Cornel Berndt just published a great Tech Byte on exactly this: how do you train a model to spot military vehicles when real images are scarce, sensitive, and expensive to collect? His answer is clever, but conditional, and what surprised him most, is that this isn't a Defence problem. Medical imaging, industrial inspection, remote sensing, no matter the industry, data scarcity is everyone's problem. Read the full article here 👉 https://lnkd.in/e8c_EunS We're growing the team. If you build AI-driven systems and want to work on problems like this, let's talk. 📩 #neverstopdeveloping #itjobs #softwaredevelopment #hiring Systematic Julian Berndt