01
Document Intelligence Systems
AI pipelines that extract structured information from large collections of documents, turning unstructured archives into searchable, analyzable knowledge.
AI & Data Engineering Studio · Portugal
Moonsteps builds AI systems that transform messy, fragmented data into structured knowledge — for investigative organizations, research institutions, and teams that need solutions beyond what off-the-shelf tools can offer.
Modern organizations generate enormous volumes of unstructured data — documents, filings, images, research archives. Extracting meaningful knowledge from these sources is difficult and often requires custom engineering.
Off-the-shelf AI tools are not enough.
AI pipelines that extract structured information from large collections of documents, turning unstructured archives into searchable, analyzable knowledge.
Systems that automatically identify entities and relationships across datasets, constructing searchable knowledge graphs from fragmented sources.
End-to-end pipelines integrating machine learning with production infrastructure — from document ingestion to semantic search and automated classification at scale.
Investigative Tools
A computer vision system capable of identifying passports and identity documents across large investigative datasets — enabling journalists and researchers to surface relevant materials from massive document archives.
Corporate Intelligence
A system that analyzes corporate filings to identify subsidiaries and ownership relationships, automatically constructing maps of corporate structures from unstructured regulatory documents.
Scientific Computing
Infrastructure and ML workflows for processing and analyzing large microscopy datasets — combining research-grade image analysis with production-ready data engineering.
Production systems built with modern AI and data engineering tools — from research prototype through to deployed infrastructure.
Moonsteps was founded by Nuno Loureiro, a physicist and neuroscientist working at the intersection of AI, data engineering, and investigative analysis.
Projects combine research-grade machine learning with production-ready data systems — built for organizations that require solutions designed for their specific domain complexity.
Moonsteps works best where data is messy, fragmented, or unstructured — where existing tools fall short and custom engineering is the only path forward.
Moonsteps collaborates with research institutions, investigative organizations, and technology teams that need custom AI solutions — for problems where existing tools fall short.
build@moonsteps.tech