Senior AI researcher and data scientist working across machine learning, numerical modeling, scientific AI workflows, and reproducible research tooling.
My work is centered on building practical AI and data-science systems for scientific problems: from uncertainty-aware ML pipelines and physics-informed neural networks to phase-field modeling, lightweight educational demos, and tools for improving research workflows.
- Machine learning and data science for scientific applications
- Uncertainty-aware modeling and reproducible ML pipelines
- Physics-informed neural networks and phase-field simulations
- Multi-phase-field and KKS-type microstructure modeling
- AI-assisted scientific workflows and manuscript quality control
- Lightweight educational demos for ML, RL, and diffusion concepts
- Applied analytics projects, including sports and engineering use cases
I develop and maintain projects around physics-informed neural networks, phase-field modeling, and ML-assisted numerical simulation. This includes work on Allen–Cahn/Cahn–Hilliard-type problems, multi-phase-field interface dynamics, and KKS-style microstructure modeling for materials systems.
Key repositories:
-
PINNs_MPF--a-Physics-Informed-Neural-Network-for-Multi-Phase-Field-problems
Physics-informed neural-network framework for multi-phase-field interface problems. -
Extended-Kim-Kim-Suzuki-KKS--Phase-Field-model-for-prediction-of-Materials-Properties
Extended KKS phase-field modeling for microstructure evolution and materials-property prediction. -
Physics-Informed-Neural-Networks-for-Allen-Cahn-equations
PINN implementation for solving Allen–Cahn equations.
Human-in-the-loop quality control for AI-assisted scientific manuscript preparation.
manuscript-qc helps researchers use AI during manuscript preparation while maintaining control over claims, citations, numerical consistency, figures, reviewer risks, and publisher-policy compliance.
I also build small, inspectable educational projects such as microRL and microDiffusion, designed to make core ML ideas easier to visualize, test, and understand without heavy infrastructure.
| Repository | Focus |
|---|---|
| PINNs_MPF--a-Physics-Informed-Neural-Network-for-Multi-Phase-Field-problems | PINNs for multi-phase-field interface dynamics |
| Extended-Kim-Kim-Suzuki-KKS--Phase-Field-model-for-prediction-of-Materials-Properties | Extended KKS phase-field modeling for microstructure and materials-property prediction |
| Physics-Informed-Neural-Networks-for-Allen-Cahn-equations | PINN implementation for Allen–Cahn equations |
| MachineLearning_UncertaintyAware_Hydrothermal_Pipeline | ML pipeline with uncertainty quantification for hydrothermal processing |
| scientific-writing-skills | Human-in-the-loop QC framework for AI-assisted scientific manuscript preparation |
| microRL | Small browser-based reinforcement learning visualization |
| microDiffusion | Minimal diffusion-model visual lab |
| Omdena-UEFAEURO2024 | Sports analytics and machine learning for football data |
| Room-Booking-System-Cpp | C++ room-booking and interval-management project |
I am particularly interested in tools that make scientific AI more reliable, inspectable, and useful in real research workflows. This includes physics-informed learning, phase-field modeling, model uncertainty, numerical reproducibility, manuscript auditability, and practical human-in-the-loop systems.
Open to collaboration on scientific AI workflows, data science, physics-informed modeling, phase-field simulation, uncertainty-aware ML, and reproducible research tooling.