Dhurba Baral
MS in Computer Science (Thesis) · University of Cincinnati
Meller Lab, College of Medicine
Cincinnati, OH, USA
I am a graduate student in Computer Science at the University of Cincinnati, where I work with Prof. Jarek Meller (Division of Biomedical Informatics, Cincinnati Children's Hospital Medical Center) on machine-learning methods for computational drug discovery.
My research involves representation-learning approaches to virtual screening and drug–target interaction prediction. It includes contrastive learning for protein-family-aware retrieval of targets for drugs, and protein-small molecules interaction prediction using interpretable deep learning methods. I also work in pharmacogenomics, modeling genotype–phenotype relationships from SNP data to predict adverse drug response. Before graduate school I spent two years as a machine-learning engineer at TAI, Inc., building LLM systems and production ML infrastructure.
Research Interests
- Using machine learning for drug-target prediction and virtual screening
- Representation learning for proteins and its structure, domain and function
- Statistical and machine-learning methods for genotype–phenotype association
Publications & Presentations
-
July 2026
Benchmarking the Generalization of Chemical Language Models
Poster, Intelligent Systems for Molecular Biology (ISMB 2026), MLCSB track, Washington, DC.