Advancing AI for real-world, on-device applications.

• Developing VLM applications for edge devices to address practical challenges.
• Working on model compression techniques to enhance AI efficiency and deployment.

Developed AI models for drug discovery to advance healthcare.

• Designed a representation learning framework for molecules of various scales, embedding small molecules to proteins in a unified space.
• Built a controllable molecular generation model using diffusion models, conditioned on protein properties.
• Identified and addressed modality imbalance in molecular interaction prediction via a multimodal-mixing approach.

• Managed and visualized SNU’s greenhouse gas and energy data using R.
• Improved data workflow and established visual baselines to track energy trends and detect anomalies.
• Gained hands-on experience with real-world data management.