Distributed & Federated Learning • Split Learning • AIoT • Edge Computing
With a PhD in Computer Engineering from Jeju National University and a career spanning academia and research, I focus on distributed machine learning and edge computing. I have led projects in split federated learning, mentored PhD students, built digital-twin platforms for FL, and explored NAS for resource-constrained devices, contributing to a body of work with a cumulative impact factor of 220+.
STINT project on Federated Learning for resource-constrained devices.
Project participation, reporting, paper writing, and PhD mentoring.
Project execution and publications.
Led programming labs in Java, Python, and C#; supervised industry-style student projects.
Adaptive client–server partitioning to reduce latency and energy for low-power edge clients.
Airborne relays for topology-aware DFL to improve connectivity, robustness, and coverage.
Mitigating resource heterogeneity and distribution shifts via split learning across cells.
Automated architecture search tailored to diverse edge capabilities and constraints.
Learning communication structures that reduce latency and improve convergence.
Remote monitoring/control of DFL clients with real-time observability and fault handling.
Hierarchical personalization for smart-home thermal comfort in heterogeneous settings.
DT framework to supervise, validate, and control centralized FL clients at scale.
Client/model selection via sampled validation to resist poisoning and low-quality updates.
Edge AI models for predictive control of indoor environmental parameters; Research Assistant role.
Geological ML to identify soft layers and estimate water levels for lower-risk drilling.
Raspberry Pi + TensorFlow forecasts with Differential Evolution to minimize energy costs.
Office # MIT.D.224, Dept. of Computing Science, Umeå University