Paul Scotti, Ph.D. is Co-Founder and CTO of Sophont, a medical AI startup building open multimodal foundation models that can reason across diverse clinical data types including pathology slides, brain scans, clinical notes, and lab results. He also serves as a Visiting Research Scientist at the Princeton Neuroscience Institute, leading projects in real-time fMRI-to-image reconstruction and brain foundation models.

Paul earned a BA in Psychology (Magna cum laude) from The George Washington University and a PhD in Cognitive Neuroscience from The Ohio State University, followed by postdoctoral research at Princeton University. His academic training forged a deep expertise in computational neuroscience and neuroimaging that spans over a decade.

Paul is the creator of MindEye, the leading pipeline for reconstructing visual experiences from fMRI brain scans, with research published at NeurIPS, ICML, CVPR, and Nature Biomedical Engineering. He is a strong advocate for open science, actively leading and mentoring international volunteer communities in NeuroAI research.

Before co-founding Sophont, Paul served as Head of Neuroimaging and AI at Stability AI, where he led the MedARC Neuroimaging and AI Lab and produced first-author papers at NeurIPS and ICML. His work bridging academic neuroscience and scalable AI has been featured in the New York Times and cited in a US Senate hearing on AI and intellectual property.

Sophont raised a $9.22M seed round in September 2025 led by Kindred Ventures, with participation from Upfront Ventures, Delphi Ventures, Jeff Dean of Google DeepMind, and Clement Delangue of Hugging Face. The company collaborates with international medical centers and academic institutions to publish open-source medical AI research and build infrastructure for med-tech and pharmaceutical R&D teams.