edgepiityped character spans
taxonomy 2026.11 GitHub
Headshot of Raphael Mansuy, author of edgepii.

About the author

Raphael Mansuy

Author of edgepii · data engineering and AI strategist

Raphael has more than 20 years of global technology leadership experience. He works on democratising data management and AI, and helps organisations build scalable data platforms. His work covers resilient data infrastructure, machine learning models, and coaching teams on practice. He leads a startup studio in Hong Kong and writes about AI and data strategy.

Why this library exists. A PII tool usually answers one question: is this document sensitive? That answer is too coarse for a support ticket, a log line, or a prompt, and it gives no proof of what left the system. The smallest honest fix is a typed range of characters on the caller's own string, with one checkable equation behind every row. So he built it, and then measured it on two public corpora rather than trusting a model card.

Connect on LinkedIn GitHub profile Repository

Recognition

Favikon ranks Raphael at 12 in Top 20 AI Experts in France in 2026, with a Favikon Authority Score of 7,012 points. The entry cites his 20 years of data and AI leadership, his work on scalable data platforms, and the startup studio he leads in Hong Kong.

Source: Favikon, Top 20 AI Experts in France in 2026.

Cover of The Definitive Guide to Data Integration, with Raphael Mansuy as a co-author.

The Definitive Guide to Data Integration

Packt, 2024

Unlock the power of data integration to efficiently manage, transform, and analyze data. Raphael is a co-author with Pierre-Yves Bonnefoy, Emeric Chaize, Mehdi Tazi, and Stephane Heckel.

Read it on O’Reilly

Other projects by Raphael

EdgeQuake knowledge graph view: 200 nodes and 840 edges over an AI research corpus.

EdgeQuake

Graph-RAG in Rust: one binary, one Postgres, any LLM. It reads PDFs and text, builds a knowledge graph in PostgreSQL, and answers questions with sources.

EdgeCrab illustration: a comic battle between EdgeCrab and OpenClaw.

EdgeCrab

A personal and coding SuperAgent in one native Rust binary, with a terminal UI and a ReAct tool loop. It supports 24 LLM providers and 17 messaging gateways.

EdgeWhisper banner: voice dictation, on-device AI for macOS.

EdgeWhisper

Voice dictation for macOS, powered by Voxtral from Mistral AI. It runs fully on the device, with no cloud and no data collection.

Handling PII in a pipeline? Tell Raphael what you are building.

Connect on LinkedIn