Summary
Matteo Omenetti is a research engineer at IBM Research in Zurich with seven years of experience building multi-modal and document-understanding systems after earning a Master's in Computer Science from ETH Zurich. He leads development on Docling, a widely used Linux Foundation-backed Python library for robustly parsing unstructured documents, and co-developed SmolDocling, an ultra-compact VLM that topped Hugging Face trends and pushed state-of-the-art results under extreme size constraints. His work spans foundation models for code and formula extraction, large-scale open datasets (e.g., SynthCodeNet, SynthFormulaNet, SynthChartNet), and production-scale inference on hundreds of millions of documents via Ray. Matteo combines strong research rigor—publishing and open-sourcing models and datasets—with practical engineering delivery, having built high-coverage data-management tooling and achieved SOTA results like 97% BLEU on Im2Latex. Known for squeezing maximal performance from tiny models, he brings a rare mix of ML research, software engineering, and large-scale systems expertise.
7 years of coding experience
1 year of employment as a software developer
High School, High School at Calabasas High School
Master's degree, Computer Science, Master's degree, Computer Science at Eidgenössische Technische Hochschule Zürich
High School, High School at Liceo Scientifico Savoia Benincasa Ancona
Bachelor's degree, Computer Science, 9.36/10, Bachelor's degree, Computer Science, 9.36/10 at USI Università della Svizzera italiana
Italian, English, German