About

I’m Jonas Vanbuel, a full-stack AI engineer based in Antwerp, specialising in custom RAG systems and LLM-powered interfaces grounded in your data — not the open web. You bring the domain expertise; I build the infrastructure that makes your knowledge base queryable and actionable. The result: untapped data turned into clear, decisive intelligence.

As a founder-engineer, I shipped Archibald from zero to production in under three months — an AI art advisor delivered entirely via email. No app, no dashboard: email a question, get multi-year market analysis in seconds. Under the hood: a hybrid text-to-SQL + vector search architecture over 8M+ auction records and 50k+ unstructured documents, cutting research time by 80% and hallucinations by 90%. From data ingestion to cloud deployment, I own the full stack.

Before that, I spent several years in the art market — consultancy and sales, moving between New York, London, and Antwerp. Along the way, I collaborated with All Ears in London, building AI-generated audio content with local TTS and voice cloning (Chatterbox) for clients including Tate Modern and Goldsmiths University.

How I Work

  1. Discovery — Mapping your data landscape and identifying the highest-value opportunities.
  2. Pipeline Design — Ingesting, cleaning, and structuring your data for AI retrieval.
  3. RAG Build — A custom AI interface grounded in your data, tailored to your domain.
  4. Deployment — Production-ready, in the cloud or on-premise, on your terms.
  5. Iteration — Continuous refinement as your data grows and needs evolve.

Tech Stack

  • AI / LLMs: OpenAI, Anthropic, Groq, Ollama, LangChain, LangGraph
  • Languages & Data: Python, pandas, FastAPI, PostgreSQL, Supabase
  • Vector & Search: Chroma, Pinecone, semantic search
  • Full-Stack: TypeScript, React, Next.js, Node.js
  • Infrastructure: Docker, AWS, GitHub Actions, OpenTofu

Working on a data or AI problem? Let’s talk — or connect on LinkedIn. Happy to think through your project with you.