Summary
Jacek Golebiowski is a machine learning scientist and entrepreneur with 11 years of experience building practical, research-driven ML systems and co-founding Distil Labs to make training small language models as easy as prompting. He led teams at Amazon/AWS developing LLM-based assistants, tool-using systems, model-and-human-in-loop evaluation, and sample-efficient hyperparameter optimization (core developer of the SyneTune HPO library). His work spans production search features—fairness-aware ranking and diversification—to compact specialized models that match larger LLM accuracy for specific safety and OOS tasks. Trained as a computational materials scientist (Imperial College PhD researcher), he blends physics-grade numerical methods with modern ML, bringing unusually strong HPC and algorithmic rigor to applied NLP. Based in Berlin, he combines startup grit with deep research chops and a track record of shipping science-backed solutions at scale.
11 years of coding experience
9 years of employment as a software developer
Master’s Degree, Theory and simulation of materials, Distincition, Master’s Degree, Theory and simulation of materials, Distincition at Imperial College London
Bachelor's Degree, Theoretical and Experimental Physics, 5 (Bardzo dobry), Bachelor's Degree, Theoretical and Experimental Physics, 5 (Bardzo dobry) at University of Warsaw
English, Polish, French, German