Larry Tesler's theorem says AI is whatever hasn't been done yet. We work on those problems for power utilities, public institutions and critical infrastructure, and we build custom AI, data science and semantic systems that run in production.
Get in touch
Customised AI and engineering solutions built around your business context, from applied AI and semantic data integration to the enterprise applications that put them to work.
Deep specialisation in the two environments where getting it wrong is not an option.
Grid-scale AI and data science for electrical and industrial infrastructure, from parsing COMTRADE records and Fourier-based fault detection to unified data models (CDM/CGMES). We build resilient backends that carry critical operational load every day.
Automated data pipelines for public-sector and administrative systems, with semantic integration, knowledge graphs, and document intelligence at scale. This includes 3M+ tender documents analyzed for the EU Commission.
Every engagement follows the same path, from a complex problem to a system that carries operational load.
We start inside your domain, not our toolkit. Working with your engineers and data, we map the real problem, including the systems, the constraints and the edge cases, and model it as ontologies and common data models (CDM/CGMES, knowledge graphs). The domain becomes the backbone the whole system is built on.
Before any model is trained or line of code is written, we define precise behaviour: expected outputs, edge cases, operational constraints, and how success is measured. A clear specification is what separates a system built for production from one built for a demo.
We build against the spec. That covers models, semantic data layers, backends, and the APIs that connect them, integrated directly into your existing workflows and infrastructure. We do not build isolated prototypes. The software is designed to live where your teams already work.
We test against the specification under real operational conditions, not ideal ones. We check performance under load, failure modes and data quality, and harden the system until it holds up against messy, real-world enterprise data every day.
We deploy into production and stay with it. Monitoring, iteration, and ongoing support keep the system resilient as data, load, and requirements evolve over time.
Proof of concept for innovative AI-based virtual assistants for industrial and offline applications
Lyra is an end-to-end, AI-powered chatbot orchestration platform engineered to function entirely offline on resource-constrained industrial devices. At its core, the system leverages a Small Language Model (SLM) integrated with a local Retrieval-Augmented Generation (RAG) mechanism. This allows the assistant to distill information from hundreds of technical manuals to provide instant, expert guidance for equipment maintenance and troubleshooting without requiring an internet connection.
Expected results
The project aims to establish a new paradigm for industrial AI, enhancing operational efficiency and reducing equipment downtime through localized intelligence. It will result in a low-level prototype (TRL 4) validated through rigorous testing in simulated industrial scenarios. By proving the innovative concept of knowledge distillation on low-power devices, Lyra secures the technical foundation for a secure, private, and high-performance alternative to traditional cloud-dependent assistants.
Development of an innovative software solution for automated reporting (VizTool) and services for the implementation of artificial intelligence in supercomputing (Super AI)
VizTool is designed for automated retrieval of data collected and stored in digital repositories, enabling users to quickly, easily, and comprehensibly group data into visual elements and reports in a short period of time. Super AI implements neural network models for processing and analyzing larger data sets instead of traditional algorithms for statistical data processing. The project implementation includes technology demonstration activities and the establishment of production for the initial market introduction. Training of employees for specific skills of the start-up company will also be conducted.
Expected results
The project is aimed at creating the conditions for the growth and expansion of AI Technologies d.o.o.'s business based on the sale of their own products and services based on artificial intelligence.