OpenGenie
Repository on GitHub

About

Portrait of Raphael Mansuy, author of OpenGenie.

About the author

Raphael Mansuy

Author of OpenGenie · 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 data platforms that survive contact with production. His work covers resilient data infrastructure, machine learning systems, and coaching teams on practice. He leads a startup studio in Hong Kong and writes about AI and data strategy.

Why OpenGenie exists. Databricks Genie Code is the best coding experience for data engineering on the market, and the reason is context: the agent sees Unity Catalog, the warehouses, the notebooks and the jobs, so it works from your data rather than from a guess. What it does not let you do is bring your own model, and its cost is metered by a product you do not configure. Raphael first measured whether that matters — eleven graded tasks in a live workspace, every run published — and found the model was the largest lever. Then he built the application the measurements implied: a Databricks App with the same kind of context, where the model is a setting, writes are fenced to a folder you name, and a claim of “done” is checked against what the agent actually ran.

Recognition

Favikon lists Raphael among the Top 20 AI Experts in France, citing his data and AI leadership, his work on scalable data platforms, and the startup studio he leads in Hong Kong.

Source: Favikon, Top AI Experts in France.

Cover of The Definitive Guide to Data Integration, co-authored by Raphael Mansuy.

The Definitive Guide to Data Integration

Packt, 2024

Unlock the power of data integration to efficiently manage, transform, and analyze data. Co-authored with Pierre-Yves Bonnefoy, Emeric Chaize, Mehdi Tazi and Stephane Heckel.

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 a crab-like rival.

EdgeCrab

A personal and coding agent in one native Rust binary, with a terminal interface and a reactive tool loop. It speaks to many model providers and 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 round trip and no data collection.

Running a Databricks estate? Tell him what Genie Code costs you, and what you would want a coding agent to prove before you trusted it.

Message Raphael on LinkedIn

This page was built from commit 7f4669e on 9 November 2026. The project is Apache-2.0 licensed.