Hyunjin Seo
Hi! I'm a Ph.D. student at KAIST, advised by Prof. Sungsoo Ahn. My research broadly focuses on applying machine learning to proteins and drug-like molecules to accelerate experimental discovery. Currently, I'm interested in developing models that predict experimentally observable outcomes, helping bridge the gap between in silico modeling and wet-lab experiments.
I'm always open to discussions and collaborations, so please feel free to reach out by email!
News
Jul 2026
TheBioCollection released on arXiv
May 2026
VibeProteinBench released on arXiv
Jan 2026
Paper accepted at ICLR '26
Aug 2025
Selected for a Research Support Grant by NRF
May 2025
MELD released on arXiv
Jan 2025
Paper accepted at ICLR '25
Dec 2024
Paper accepted at AAAI '25
Oct 2024
ReBind released on arXiv
May 2024
RoSE released on arXiv
May 2024
ICLR '24 @ Vienna, Austria
Jan 2024
Paper accepted at ICLR '24
Jul 2023
Paper accepted at ICCV '23
Mar 2023
Started my M.S. & Ph.D. journey at KAIST
Publications
TheBioCollection: Unified Pre-Training Scale LLM Corpus for Biology
Atom-level Protein Representation Learning Improves Protein Structure Prediction
VibeProteinBench: An Evaluation Benchmark for Language-interfaced Vibe Protein Design
Boltz is a Strong Baseline for Atom-level Representation Learning
Learning Flexible Forward Trajectories for Masked Molecular Diffusion
ReBind: Enhancing Ground-state Molecular Conformation Prediction via Force-Based Graph Rewiring
TheBioCollection: Unified Pre-Training Scale LLM Corpus for Biology
Atom-level Protein Representation Learning Improves Protein Structure Prediction
VibeProteinBench: An Evaluation Benchmark for Language-interfaced Vibe Protein Design
Boltz is a Strong Baseline for Atom-level Representation Learning
Learning Flexible Forward Trajectories for Masked Molecular Diffusion
ReBind: Enhancing Ground-state Molecular Conformation Prediction via Force-Based Graph Rewiring
Towards Precise Prediction Uncertainty in GNNs: Refining GNNs with Topology-grouping Strategy
Unleashing the Potential of Text-attributed Graphs: Automatic Relation Decomposition via Large Language Models
TEDDY: Trimming Edges with Degree-based Discrimination strategY
PC-Adapter: Topology-Aware Adapter for Efficient Domain Adaption on Point Clouds with Rectified Pseudo Label
Efficient Subword Segmentation for Korean Language Classification
Analysis and Utilization of Hypergraph Neural Networks
Relation-aware Pseudo-labeling for Link Prediction with Graph Neural Networks
Education
Korea Advanced Institute of Science and Technology (KAIST)
2023.02 — Present
- M.S. & Ph.D. integrated student
- Kim Jaechul Graduate School of Artificial Intelligence
Korea University
2018.02 — 2022.08
- Bachelor’s degree
- History, Artificial Intelligence (Double Major)
Experience
AI Research Intern · Trillion Labs
- Built benchmarks and a large-scale training corpus for protein design and broader biology, gaining hands-on experience in LLM mid- and post-training. Also developed an agent framework that unifies protein design tools for autonomous end-to-end workflows.
AI Research Intern · Polymerize
- Headed an independent research team focused on multi-objective molecule generation, tackling a wide range of structures like polymers and drug-like molecules.
Teaching Assistant
- Machine Learning for Artificial Intelligence (KAIST)
Research Intern · Machine Learning and Intelligence Lab (MLILAB), KAIST
- Advisor: Prof. Eunho Yang · Mentor: Joonhyung Park (Ph.D. Candidate, KAIST)
- Tackled efficient domain adaptation for point clouds, critical due to distributional shifts in real-world applications, by developing GNN-based topology-aware adapters to seamlessly adapt models from source data to target local characteristics.
Research Intern · Data Mining and Information Systems Lab (DMIS), Korea University
- Advisor: Prof. Jaewoo Kang · Mentor: Seongjun Yun (Applied Scientist, Amazon)
- Developed a pseudo-labeling approach for graph edge prediction task, assigning labels to disconnected node pairs using a hierarchical scoring mechanism.
Projects
Development of Sovereign AI for Protein Complex Structure Prediction
Ministry of Science and ICT (MSIT)
Structural Guidance for Enhancing Molecular Understanding in Large Language Models
Samsung Electronics
Molecular Property Prediction via Denoising-based Graph Neural Networks
Samsung Electronics
Predicting Food Product Satisfaction using Large Language Models
LAB-EAT
Awards
Research Support Grant
National Research Foundation of Korea (NRF)
Best Undergraduate Award
Korea Computer Congress (KCC) ·
Top 2.7%
Great Honors Award
Korea University
Services
Invited Talks
- Learning Flexible Forward Trajectories for Masked Molecular Diffusion
- LG Materials Intelligence Lab (Seoul, South Korea)
Conference Reviewer
- International Conference on Learning Representations (ICLR)
- Neural Information Processing Systems (NeurIPS)