The research partner you don't have to mention
Bespoke data science for platform accountability. White-label. Confidential. Ready to act on.
We build the systems that answer hard questions.
Monitoring & Alerts
Detection pipelines for content, behaviors, trends. Real-time alerts. You define what matters. We build the system that watches for it.
Classification Systems
Models trained on your data, validated by your people. Off-the-shelf tools miss context because they don't know your problem.
AI Integration
The model is the easy part. We build the infrastructure around it: data pipelines, workflow integration, state-of-the-art models online or locally on your infrastructure.
Your tools weren't built for your problems.
Off-the-shelf tools disappoint because they don't know your problem. We tailor.
We become part of your team, understanding your specific needs and constraints.
Point us to your data and tell us where you need results. We build everything in between.
Everything reproducible, automatable, ready for handover. Or we keep running it.
What We've Shipped
We work on sensitive projects. We don't name clients. Here's what we can tell you.
Multilingual classification system for illegal speech detection across major platforms
Digital threats monitoring products tracking coordinated campaigns and manipulation
Native-level classification including Baltic, Central European, and Slavic languages
Regulatory-grade evidence standards for sensitive harm classifications
Systems We've Built
Classification systems deployed for DSA compliance testing
Real-time monitoring infrastructure for political communication research
Algorithmic recommendation audit frameworks
Multi-platform content moderation analysis pipelines
References available under NDA for qualified inquiries.
Publications
Our public research demonstrates our domain expertise. Most of our work is confidential.
LatestLund University Psychological Defence Research Institute · December 2025
AI-Assisted Analysis of War-Related Content on Grey Zone Domains
Working Paper 2025:1
Grey zone domains aggregate violent content from messaging apps and social media, reaching hundreds of thousands of views per post. We built an AI-assisted system to analyse this content while minimising analyst exposure. About 17% of posts related to ongoing conflicts.
Read more →
