Patrick Danto

I design demanding technical architectures for the social and medico-social sector — a field where data is dense, regulated, and where generalist systems quickly reach their limits.

My background is atypical: 12 years of operational experience in the sector (pricing, budget management, CPOM, establishment management), training in quantitative economics and public policy (Sciences Po / ENSAE), and initial experience in data analysis at McKinsey Research & Insights. Today, as the founder of ConfidensIA, I develop specialized ESSMS AI bricks — and I take on freelance assignments for projects where business integration makes the difference.

What interests me: architectures where field knowledge is not an afterthought, but a design constraint from the start.

What I build and what I work on:

Sector-specific RAG systems — dense documentary corpora (CCN, jurisprudence, HAS recommendations), thematic routing, regulatory safeguards integrated into the generation chain

NLP and pseudonymization — NER models fine-tuned on medico-social corpora, identifying entity taxonomies, privacy by design pipelines before indexing or external LLM calls

ESSMS data pipelines — base enrichment (FINESS), signal detection, automated parsing and analysis of budget documents (CA, BP, EPRD/ERRD)

Framing and technical specification — translating ESSMS business challenges into actionable architectures, feasibility audits, supporting technical teams without sector expertise