Benjamin Polge

Benjamin Polge AI & Agentic AI Expert

Specialist in LLM, VLM and agentic workflows. I lead the AI section at Journal du Net and help, through my articles and analyses, IT leaders shape their AI adoption strategy.

Based in Serris (Val d'Europe), Île-de-France - Paris region

From the blog

Start with my complete guide to agentic AI

Definitions, MCP/A2A protocols, coding agents, security guardrails, and enterprise use cases: this is the strongest non-branded entry page on the site.

Read the guide

From the lab

DLPTime.com - AI in production for Disneyland Paris

A measurable AI system shipped to production: real-time wait-time predictions powered by Google's TimesFM model after an A/B shadow (MAE -11% vs baseline). Stack: Nuxt 3 SSR, bilingual PWA, PixiJS map.

Visit DLPTime.com

Interactive Demo

Try my integrations in the AI Sandbox

Ask questions to my assistant grounded in my resume, personal projects, and recent articles. Practical proof of prompt engineering and model integrations in production.

AI expertise snapshot

From newsroom strategy to hands-on builds

This portfolio mirrors my CV: AI Section Manager at Journal du Net, orchestrating LLM/VLM analyses, shipping interactive AI widgets, and combining on-field photography with data-driven reporting when useful.

AI Section Manager – Journal du Net

Lead editorial strategy for the AI channel, brief C-level stakeholders, and publish a weekly AI newsletter covering enterprise adoption, governance, and policy.

LLM/VLM benchmarks & prompt systems

Evaluate Claude, Gemini, GPT, Mistral, and open models; design prompt libraries, vibe-coding workflows, and multi-agent stacks tailored to newsroom and partner needs.

AI products + field reporting

Build Cursor/Claude Code prototypes, data visualisations, and AI-powered widgets while producing Hans Lucas photo assignments that ground each story in reality.

AI journalism & research

Recent investigations & articles