Hi, I'm Franco, CTO at a68, building fraud detection for insurers

Hi all, I'm Franco, Co-Founder and CTO of a68, based in Frankfurt. We build Fraud Catcher, an AI tool that helps insurers spot suspicious claims by checking documents, images, invoices and claim history. It combines specialized AI agents, computer vision models, classic machine learning and forensic review modules, with a supervisor agent that pulls the findings into one traceable assessment for claims teams.

My background is a PhD in applied mathematics focused on AI, with research in computer vision. I also publish small open-source tools on GitHub, mostly image annotation and speech-to-text. Right now I'm most interested in systems with loops, and in how to keep agent-based review explainable enough for claims teams to trust.

What I'd love from this community: feedback or intros from anyone who has built agent pipelines where a human has to trust and act on the output.