Research
Navigating chemical space
I develop computational and AI methods that connect structure generation to prediction, simulation, decision-making, and reproducible scientific workflows.
Small molecules → polymers → crystals → MOFs
My current expertise is strongest in small-molecule chemical-space exploration, quantum chemistry, property prediction, and workflow automation. Polymers, crystals, and metal–organic frameworks are intentional expanding directions—not claims of completed expertise.
LEGO-based generation
Reusable blocks, controlled assembly
The strategy begins with chemically meaningful fragments rather than arbitrary atoms. Connection rules assemble them into candidate molecules or structures, property models screen the generated space, and quantum chemistry refines selected candidates. This makes the design logic traceable and adaptable across related chemical domains.
- RepresentDefine fragments and connection sites.
- GenerateExplore combinations under chemical rules.
- PrioritizeUse predicted properties and synthesizability.
- ValidateApply quantum and atomistic calculations.
Current expertise
Established foundation
- Chemical-space generation and exploration
- Quantum chemistry and DFT workflows
- ML-driven molecular property prediction
- High-throughput screening and HPC automation
- Scientific Python, web tools, and agents
Emerging directions
Expanding capability
- Polymer representations and generative workflows
- Crystal structure exploration and ML potentials
- MOF representation, generation, and screening
- Active learning and autonomous optimization
- Literature mining and scientific agents