People

Ali came to crop science from fluid mechanics. He studied mechanical engineering at Sharif University of Technology, modeled cardiovascular and respiratory flows at Northern Arizona University, and completed a Ph.D. with Aaron Towne at the University of Michigan, where he built RSVD-Δt, a parallel algorithm that computes resolvent modes with linear scalability.

He now builds DNA language models for plant genomes. His CASCADE framework trains a cell-type-resolved sequence-to-expression model on a single-cell soybean atlas, then reads the model’s own attributions against a position-specific null to recover the promoter sequence driving expression — turning a trained model into a map of candidate regulatory elements for crop design.

Juan Pablo Muñoz Díaz completed his Ph.D. under the supervision of Prof. Jesper Tegnér (KAUST), co-supervised by Dr. Narsis Kiani (Karolinska Institutet). His research merges machine learning and bifurcation theory to uncover hidden variables in nonlinear biological systems. He has conducted collaborative research at the University of Cambridge and presented at leading international conferences, including the International Conference on Systems Biology and the SIAM Conference on Dynamical Systems. He is now working on Mechanistic AI for modeling biological and crop growth, linking genes to intermediate and final phenotypes.

Olatunde builds large-scale genotype-by-environment (GxE) models for peanut, corn, and cotton. His work draws on open-source remote sensing, soil, and weather datasets, assembling them into a common spatiotemporal record of the growing environment, and then links that record to crop performance using spatiotemporal graph neural networks and deep operator networks.

The aim is a modeling framework that predicts how a given genotype will perform across a landscape and across a season, rather than at a single site at a single point in time, and that does so from data already available in the public domain.

Roth builds agents for bioinformatics, applied to plant biology and breeding. His work packages the analyses that normally demand a specialist — genome assembly and annotation, comparative genomics, population-scale sequence analysis — into agentic systems that can plan and carry out multi-step workflows on their own, so that biologists and breeders can put questions directly to their sequence data.

Ashmita is a Ph.D. student in the Josh Clevenger Lab at the HudsonAlpha Institute for Biotechnology, and an affiliate member of the Computational Agriculture Lab. She works across high-resolution genomic resources, modeling methods, and phenotyping technologies to improve breeding in peanut and blueberry. Her research pairs dense genomic data with new phenotyping platforms and develops the statistical and machine learning methods needed to connect the two, with the goal of shortening selection cycles and making breeding decisions in these crops more predictable.

Phong builds phenotypic pipelines that detect and quantify disease in peanut, turning field and greenhouse imagery into quantitative measures of disease severity and spread. He also designs components for self-driving laboratories, developing the automation that makes gene editing experiments faster, cheaper, and more repeatable.

David builds phenotypic models that detect and interpret drought stress in plants. His work focuses on recognizing the visible signatures of water limitation early, and on connecting those signatures back to the underlying physiological response, so that stress can be measured and understood before yield loss becomes irreversible.