AI studioMaribor, Slovenia
AI can't think. It's powerful math.
We build production AI systems that carry real load, every day, for insurers, defense programmes and the public sector. We will tell you what the technology can do. We will also tell you what it cannot. The second half is the part most people skip.
No pitch decks. No sales calls. Just a conversation.
// trusted by
About
Serious engineering. Serious systems.
CODE-RED is a specialized AI engineering team based in Slovenia. We build systems that run in production, process millions of documents, and operate in environments where failure has real consequences.
From autonomous document processing for the largest insurance company in the region to advanced AI systems for national defense, we work where it matters.
We have been doing this for over 8 years, before AI was a boardroom buzzword. When a client trusts us with their most critical workflows, we deliver.
No demos that go nowhere. No proofs of concept that collect dust. Working software, deployed and running.
Capabilities
What we build.
Autonomous document processing, workflow automation, intelligent decision systems with human-in-the-loop oversight. Built for scale, deployed for reliability.
Real-time object detection, image classification, visual inspection systems. Custom-trained models for specific domains, from manufacturing to gaming.
Native and cross-platform apps with on-device AI inference. From health tech to gaming to education. AI that runs locally, responds instantly.
Advanced AI solutions for military and defense applications. Systems built for environments where reliability is non-negotiable and downtime is not an option.
Taking cutting-edge AI capabilities and integrating them into existing business workflows. Making AI practical, explainable, and deployable.
Not everything needs AI. We build custom web apps, internal tools, APIs, and full-stack platforms, tailored to your exact requirements. Clean architecture, built to last.
Case Studies
Deployed. Running. Proven.
Advanced AI for the Slovenian Armed Forces. Anomaly detection and forecasting across deployable energy infrastructure, running at the tactical edge.
Built to defense interoperability standards: MIL-STD-3071 over a DDS data layer, with our own AI components following ISO/IEC 22989 and ISO/IEC 23053.
DETAILS CLASSIFIED
A family of document AI systems running inside the largest insurer in the region. Signature verification, identity document extraction and document type classification on one side. Invoice reconciliation, health documentation, fraud checks and claims support on the other.
Every field the systems extract flows into a single searchable store, so six years of paperwork behaves like a database rather than an archive. Questions that once meant opening documents one at a time now come back answered, aggregated and charted.
Six years in production, entirely on-premise, every model trained on their own data.
Truck height analysis and monitoring for European transport operations. AI detection of height-related hazards with real-time alerting, preventing incidents before they happen. Includes a complete transport tracking platform.
Full stack ownership: low-level networking and device firmware through cloud infrastructure to server systems and application development. End to end, single team.
AI-powered signature detection and verification system. Automatically locates, extracts, and validates signatures across document types, from contracts to insurance claims. Detects invalid and forged signatures. Built for high-volume document workflows where manual review cannot scale.
Mobile app using on-device computer vision for Overwatch. Custom YOLO model detecting 50+ heroes in real-time. Zero latency, zero server dependency, pure on-device AI inference.
overhelper.com →AI-powered language learning app for children, built for the European Centre for Modern Languages (Council of Europe). Interactive, engaging educational AI designed for young learners.
How We Work
From zero to production.
We start by understanding your domain, workflows and pain points. Not just requirements gathering, but real understanding of how your business operates and where AI creates genuine value.
We design and build the system. AI models trained on your data, integrated into your infrastructure, tested against your edge cases. No shortcuts, no black boxes.
We don't hand over a prototype and leave. We deploy to production, monitor performance, and make sure the system holds up under real load with real data.
Systems evolve. We maintain, retrain and improve. Your AI gets better over time, adapting to new patterns and edge cases as your business grows.
FAQ
Frequently asked.
- How long does a typical project take?
- It depends on the scope. A focused AI system, such as a document classifier, anomaly detector or workflow automation, typically takes 6 to 12 weeks from kickoff to production. Larger platform builds run 3 to 6 months. We'll give you a realistic timeline after the first deep dive, not before.
- Do you work with our existing infrastructure?
- Yes. We integrate into what you already have: your cloud, your databases, your CI/CD pipelines. We don't ask you to rip and replace. If you run on-premise or air-gapped, we've done that too.
- What happens after deployment?
- We don't disappear. Every system we deploy includes monitoring and maintenance. Models drift, data changes, business needs evolve. We retrain, optimize and adapt the system over time. You're not locked into a contract. You stay because it works.
- Can you work in classified or regulated environments?
- Yes. We build systems for the Slovenian Ministry of Defense and have experience with classified-grade deployments. We understand security clearances, air-gapped networks and compliance requirements.
- What if we're not sure AI is the right solution?
- That's what the first conversation is for. We'll tell you honestly whether AI makes sense for your problem. Sometimes the answer is a simpler system, a better data pipeline, or a process change. We'd rather say no than build something that doesn't deliver value.
- How does pricing work?
- We scope every project individually after the initial deep dive. No surprise invoices, no runaway budgets. You'll get a clear proposal with defined deliverables and milestones before any work begins. For ongoing systems, we offer maintenance agreements tailored to your needs.
Contact
Let's build something that matters.
Tell us about your challenge. We'll tell you if we can solve it.
- info@code-red.si
- phone
- +386 70 449 788
- office
- Beloruska ulica 7, Maribor
- HQ
- Lackova cesta 15, Maribor
- socials
- LinkedIn · Facebook · Instagram