Thomas van Dongen is Co-Founder of Semble, a fast and accurate open-source code search library built for AI agents, developed under MinishLab.

Thomas serves as Head of AI Engineering at Springer Nature, where he focuses on recommender systems and information retrieval pipelines for scientific publishing.

Thomas is a founding member of MinishLab and a passionate advocate for small, efficient, eco-friendly ML models, particularly static word embeddings, believing they remain highly relevant in the age of transformers.

Alongside co-founder Stephan Tulkens, Thomas has built a suite of open-source ML packages under MinishLab including Model2Vec, SemHash, Vicinity, Tokenlearn, and Pyversity, earning recognition in the ML community.

Semble delivers ~200x faster indexing and ~10x faster queries than code-specialized transformers at 99% retrieval quality, using ~98% fewer tokens than grep+read, running entirely on CPU with no external dependencies.