Bob van Luijt

Bob van Luijt is co-founder and CEO of Weaviate, the open-source vector database built for storing and searching machine-learning embeddings. A Dutch technologist who started companies as a teenager and studied music seriously, he has spent years arguing that language and software evolve together — and that AI apps need a database designed for vectors from scratch.

Teenage software and a music detour

He started an internet business at fifteen, later studied music at ArtEZ and Berklee, and completed Harvard Business School's Program for Management Excellence. Consultancy work under Kubrickology and a TEDx talk on digital technology through the lens of language kept him between code and composition. Open-source business models clicked when he heard Sam Ramji talk about them; he has said building infrastructure software on open-source principles felt like the natural fit.

Inventing Weaviate before the category had a name

He began Weaviate in 2016 as an open-source search engine for embeddings. On Data Science Dojo he recalled early days when he still called it a knowledge graph on a Google Cloud interview "because the term vector database didn't exist yet." The community asked for vector search; rather than wrapping Facebook's libraries the way Lucene-era search tools did, his team's bet was to "build something from scratch that is really like, that's a product."

Open source as transparency for heavy lifting

On James Le's Datacast he defined Weaviate plainly: "Weaviate is a vector search engine, which means it's a database of vector embeddings." Open source, he argued, builds transparency when you are doing heavy engineering in a new niche: "Being able to openly talk about that and show what we're working on is something that open source enables us to do." The commercial company sells cloud and hybrid deployments around the same core.

On AI-first databases

He talks about vector search as a distinct database niche — not a warehouse or time-series store with embeddings bolted on — and about hybrid search, filtering, and model integrations as the product surface teams need. Music trained his ear for structure; databases became the instrument. The through-line is infrastructure for the embedding era that stays open enough for builders to inspect and run themselves.