Medius

UN-awarded AI & software engineering company.

● 350+ projects delivered since 2002, 100% client return rate
● Trusted by Volvo, SKF, Novartis, Hisense
● Focus: tailor-made, data-driven solutions, not off-the-shelf products
● Gamayun AI platform: real-time machine learning for fraud detection, predictive
maintenance, telecom optimisation, smart cities & more
● Awards: GoDigital 2023 (Best AI Project), Deloitte Impact Star 2024, GZS best
2025 SME (IT company) in Slovenia

Since 2002, we have completed over 350 data-driven projects across various industries and
are proud to maintain a 100% client retention rate. We serve customers in nine industries,
including manufacturing, pharmaceuticals, transportation, telecommunications, energy, and
insurance. Our clients include internationally renowned companies such as Volvo Group,
SKF, Novartis, Nomago, Petrol, HSE and Hisense. For each customer, we help transform
data into a competitive advantage.

We build on proven, production-tested open-source technologies to deliver solutions that
integrate machine learning at their core. Our stack leverages methodologies such as ML,
open APIs, data streams, semantic interoperability, data lakes, CI/CD, containerization, and
cloud orchestration. This gives clients and project teams a head start—freeing them to focus
on use-case analysis, data consolidation, feature engineering, and model development
rather than technology selection or tool hunting.

Based on years of experience, Medius has developed a real-time streaming machine
learning platform – Gamayun AI – which integrates best-of-breed technologies, including
Kubernetes (microservices orchestration), Kafka (real-time data streaming), Elastic,
OpenSearch, Cloudera, Hadoop (data lakes), PyTorch, TensorFlow, Spark, NLP
(machine learning model development), and Maven, Nexus, GitLab (CI/CD). Gamayun AI is
use-case agnostic, as demonstrated by references from diverse clients and domains.

For example, Gamayun is used for fraud detection in insurance, predicting failures of data
sources in public administration, identifying critical paths in interoperable systems, predicting
bottlenecks, errors, and application failures in telecommunications, forecasting pipe
leakages in smart city water distribution, segmenting lottery players to ensure
responsible gaming, named entity recognition (NER) in public administration, and
forecasting machine downtime before it occurs