Selected case studies
AI projects we built entirely within our clients' own infrastructure - fully on-premises, on open models with strong Polish-language understanding. For each deployment we built a complete AI platform in the class of Red Hat OpenShift AI: from model development and serving to GPU-optimized inference, backend, frontend, and integrations.
Semantic contract search in a RAG system
An advanced RAG system integrated with a SharePoint contract repository and vector database. Users ask questions in natural language and receive accurate answers grounded in document content.
PL
Polish-language understanding
Full Case StudyText2SQL - querying databases in natural language
A Text2SQL system that lets users query Oracle and Teradata databases in natural language. The solution runs locally on open-source models tuned for Polish.
NL / SQL
Natural-language queries
Oracle
Supported databases
Teradata
Supported databases
Full Case StudyDocument classification and data extraction
A solution for classifying and extracting data from unstructured documents: PDFs, DOCX files, and scans. The client analyzes documents instantly and makes decisions without manual re-keying.
LLM
Classification & extraction
PL
Polish-language understanding
Full Case Study