AI agent disclosure
Last updated: April 2026
The chat on this site is powered by an AI agent, not a human. This page explains how it works, what it can and cannot do, and how to reach a human.
What it is
A LangGraph agent orchestrating multiple large language models (Anthropic Claude, OpenAI GPT, Google Gemini) routed via litellm based on the question type. It has access to a small set of tools to search Vincent's CV, score job descriptions, qualify recruiting leads, and book a meeting.
Guardrails
Following ISO 42001 principles, the agent will not lie, will not leak personal data, will not quote a daily rate (TJM), and will refuse jailbreak attempts. It always answers in the language you wrote in.
What it can get wrong
Like any LLM, it can hallucinate. If it answers something that contradicts the CV or this site, trust the site, not the agent. When in doubt, ask for an introduction or book a meeting with Vincent directly.
Transparency
Conversations are traced through LangFuse for evaluation and improvement. We do not use your messages to train models. Tool calls are visible in the chat (e.g. "Searching Vincent's experience…").
Reach a human
At any point you can ask the agent to schedule a meeting via Cal.com or trigger an email to Vincent. The agent never reveals his personal contact details — Vincent receives your message and replies himself.