Portrait of Mingze Dong, Ascent Fellow 2027
2027 ASCENT FELLOW

Mingze Dong

Postdoctoral Fellow
Stanford University
AI for Science (Bio)

Highlighted research

Stack: In-Context Learning of Single-Cell Biology

bioRxiv preprint, 2026

Biography

Mingze Dong is a postdoctoral fellow at Stanford University, working at the intersection of AI and biology. He received his PhD from Yale University and has conducted research at the Arc Institute, Stanford University, and Anthropic.

Combining causal representation learning, identifiability theory, high-dimensional statistics, and deep learning, he develops principled approaches to modeling complex biological systems. His work includes SIMVI, scShift, and CINEMA-OT for disentangling biological states and perturbation effects; R²MAE for understanding and improving self-supervised pretraining; and Stack, a single-cell foundation model enabling in-context learning and counterfactual prediction across diverse cellular contexts. At Anthropic, he developed self-evolving agentic frameworks for red-teaming frontier AI systems in cybersecurity and biosecurity.

His broader vision is to build generalizable, interpretable virtual cells through closed-loop integration of wet-lab experiments and AI models, enabling accurate prediction of cellular responses and accelerating biological discovery and therapeutic development.