Szczepan Baran

Biography
Szczepan Baran began his research career in human leukocyte antigen typing at Fred Hutchinson Cancer Research Center, where a patient's data was valuable precisely because the combination was rare. That same rarity is exactly what makes such data resist anonymization. Two decades later he works on both halves of that problem. At Instem he leads scientific and AI strategy across predictive modeling, translational analytics, and data-driven drug development, spanning preclinical safety through clinical delivery and data transparency. He previously led emerging technologies at Novartis and directed science at VeriSIM Life. He co-chairs a Foundation for the NIH network setting evidentiary standards for new methodologies with the U.S. FDA and EMA. His published work spans AI-driven patient stratification using omics and clinical biomarkers, translational modeling, and AI in regulatory science. His work centers on one question: how does this reach the patient faster and more safely?

---

Talk
Safe but Unusable: Where Anonymization Breaks Precision Medicine
A dataset can clear every re-identification threshold and still be useless. In small biomarker-defined cohorts that is the normal outcome, not an edge case. This panel brings simulation data on where the breakdown occurs, asks who is accountable for noticing, and tests what synthetic generation genuinely solves.

Large-Scale Data Solutions Showcase:
Instem
d-wise, an Instem company, builds regulated data infrastructure for clinical development: validated cloud environments for biostatistics and bioinformatics teams, and software and services that anonymize clinical trial data and documents for regulatory disclosure and research reuse — including the stratified cohorts hardest to release.