Luka Suhadolnik

Building bridges between experiments and digital workflows: lessons from Quipnex

While AI promises to revolutionize scientific discovery, experimental research still faces a foundational bottleneck: poorly structured data, inconsistent documentation, and slow, manual workflows. In this talk, I present Quipnex, a lightweight digital platform developed to bring structure, traceability, and automation to experimental research across disciplines. Based on real challenges encountered in academic labs, including my work at the National Institute of Chemistry in Ljubljana, I will discuss the gaps between AI-ready data pipelines and the reality of day-to-day experimental work. These include the lack of standardized data capture, difficulties comparing and reproducing results, and the absence of usable digital infrastructure to support ML workflows.

Quipnex was created as a pragmatic response to these issues. Not as a machine learning tool, but as an infrastructure layer to support reproducibility and data integration. It is used in several labs and is now evolving toward enabling automation and preparing research workflows for future AI applications.

I will conclude with reflections on what it takes to make labs AI-ready, and how tools like Quipnex can bridge the gap between experimental science and data-driven automation.