About Me
Hello! I’m Thao, a Ph.D. candidate in the Siebel School of Computing and Data Science at the University of Illinois at Urbana-Champaign (UIUC). I am fortunate to be advised by Prof. Heng Ji and am a member of the BLENDER Lab.
I am also currently working with the Molecule Maker Lab Institute (MMLI), where I build LLM-powered agents and foundation models for molecular representation, property prediction, and generation / optimization for drug design and materials science.
I’ve interned at Prescient Design, Genentech / Roche (Summer 2026), and am currently interning at Apodex.
Research interests
My research interests center on combining large language models (LLMs) with graph neural networks (GNNs) to build agentic AI systems for scientific discovery, with a focus on drug and materials design. Specifically, I am drawn to the following areas:
- LLM-powered agents for autonomous molecular design and optimization
- Molecule-native representations and modular chemical language models
- Synthesizable and fragment-based molecule generation
- Structure-guided optimization for binding specificity and protein design
- Foundation models for molecular and protein representation
- High-fidelity, interpretable molecular property prediction (e.g., solubility, functional groups)
Experience
Intern @ Apodex
Sep 2026 - Present
Working as an intern at Apodex.
Intern @ Prescient Design, Genentech / Roche
Summer 2026
Interned at Prescient Design, Genentech / Roche.
Ph.D. student @ UIUC
Aug - 2023 Present
Started Ph.D in CS at UIUC, joined BLENDER lab.
RA @ VISHC
Jan 2022 - Jun 2023
Started working at VISHC as a research assistant.
Internship @ VinBidata
Jan 2021 - Dec 2022
Started working at Medical Image Processing Department - VinBigdata JSC.
Undergrad
Sep 2016 - Dec 2020
Received BSc degree in Biomedical Engineering from HUST.
Selected Publications
[15] Thao Nguyen, Heng Ji. SpecOpt: Contact-Diff Reasoning for Agentic Molecule Optimization Toward Binding Specificity. preprint arXiv:2609.21165, 2026.
[14] Thao Nguyen, Saman Shafaei, Zhengyi Zhang, Huimin Zhao, Heng Ji. EnSol: an environment-aware graph neural network for molecular solubility prediction. preprint arXiv:2609.21151, 2026.
[13] Thao Nguyen, Jeonghwan Kim, Zhenhailong Wang, Heng Ji. Fraglingo: Molecular Design via Attachment-Aware Autoregressive Fragment Generation. preprint arXiv:2609.13519, 2026.
[12] Thao Nguyen, Heng Ji. MolLingo: Molecule-Native Representations for LLM-Powered Scientific Agents. preprint arXiv:2605.27853, 2026.
[11] Thao Nguyen, Kuan-Hao Huang, Ge Liu, Martin D. Burke, Ying Diao, Heng Ji. FARM: Enhancing Molecular Representations with Functional Group Awareness. Transactions on Machine Learning Research, 2026.
[10] Ziwen Wang, Jiajun Fan, Ruihan Guo, Thao Nguyen, Heng Ji, Ge Liu. ProteinZero: Self-Improving Protein Generation via Online Reinforcement Learning. Transactions on Machine Learning Research, 2026.
[9] Carl Edwards*, Chi Han*, Gawon Lee, Thao Nguyen, Bowen Jin, Chetan Kumar Prasad, Sara Szymkuć, Bartosz A Grzybowski, Ying Diao, Jiawei Han, Ge Liu, Hao Peng, Martin D Burke, Heng Ji. mCLM: A Function-Infused and Synthesis-Friendly Modular Chemical Language Model. In Proceedings of the Fourteenth International Conference on Learning Representations (ICLR) (Oral Presentation), 2026.
[8] Ziwen Wang, Jiajun Fan, Thao Nguyen, Heng Ji, Ge Liu. Variational Supervised Contrastive Learning. In Advances in Neural Information Processing Systems (NeurIPS), 2025.
A Little More About Me
Outside of research, I enjoy playing table tennis, traveling, photography, drawing, and graphic design.
