Nathan Amar, founder of JNS-IA

Background

Nathan Amar is an artificial intelligence engineer. Before founding JNS-IA, he held Head of Data & AI roles and led data and AI projects at the heart of organizations: from prototype to production, alongside the teams who use them every day.

JNS-IA carries that work forward in three ways: training teams in generative AI, designing AI agents and automations connected to the company's own tools, and advising leadership on AI strategy. Training programs are designed and delivered by Nathan, through certified training partners.

He is part of the Who's Human IA 2026 official selection: one of the 300 people selected in France for their ethical and human commitment to AI.

Verifiable certifications

Six of the certifications earned, with a link to the original certificate from the issuer where one exists.

• Claude Code in Action — Anthropic (AI agents) · verify the certificate
• Machine Learning Operations (MLOps) — Google Cloud (MLOps) · verify the certificate
• Data Science Professional Certificate — HarvardX (Data science) · verify the certificate
• AWS Cloud Practitioner — Amazon Web Services (Cloud)
• Machine Learning Scientist Career Track — DataCamp (Machine learning)
• Dataiku Core Designer — Dataiku (Data platform)

Public profile: LinkedIn.

Editorial method

This site has one simple goal: to help leaders and teams decide what to do with generative AI. Here is how what gets published on it is produced.

Service pages and resources

They are written by Nathan Amar, drawing on his engagements. The figures they show (years of experience, people trained) are the same everywhere, and no pricing is published: cost depends on scope, and is quoted on request.

The blog (in French)

The blog is published in French. A new article comes out every Monday. It is written with the assistance of Claude, Anthropic's model, from a web search covering the previous eight days, and published under Nathan Amar's editorial responsibility. We say so because readers have a right to know how a text was produced — and because an AI firm that hid it would be contradicting itself.

The rules every article must follow:

• primary sources first — the vendor's announcement, its documentation, the official text of a decision — and the press only as a fallback;
• no figure, price or benchmark written from memory: what the research hasn't confirmed doesn't get written;
• every announcement is dated and attributed, and a news article ends with its sources — without them, it isn't published;
• no set length: an article says what its subject calls for;
• the “last updated” date only changes when the content actually changes.

Article illustrations are generated by an image model (gpt-image-1, OpenAI); they evoke the topic and depict no real person or product.

Report an error

Spotted something inaccurate or outdated? Email nathan@jns-ia.com: the correction is made and dated in the article.