Tien-Dung Nguyen
About
I am a Senior Data Scientist at Techcombank and a Mathematics and Informatics graduate of the HUST Talent Program. My research lies at the intersection of applied mathematics, graph structure, network dynamics, and stochastic processes.
I use graphs as a relational language for studying heterogeneous, evolving systems and for building AI methods that operate under uncertainty. Alongside this research, I have five years of experience developing production machine-learning systems in banking and computer vision.
Research direction
Three questions currently guide my work:
- How can graph communities be updated efficiently as their structure changes over time?
- How can stochastic processes quantify propagation, risk, and uncertainty on networks?
- How can graph structure connect heterogeneous modalities while keeping a model’s assumptions and limitations explicit?
My current projects approach these questions through dynamic community detection, random-walk refinement, network risk modelling, and graph representation learning for noisy observations.
Selected research
Random-Walk Refinement for Dynamic Community Detection
RIVF 2025 work extending DF-Louvain with a random-walk-based refinement that can split communities after structural changes when the split improves modularity.
Graph-Based COVID-19 Risk Assessment Model
Markov-chain random-walk risk assessment across approximately 10,600 Vietnamese administrative units, developed while contributing to Vietnam’s national COVID-19 response.
Contrastive Viewpoint-Aware Shape Learning for Long-Term Person Re-Identification
WACV 2024 research on combining pose-graph shape cues, appearance, and viewpoint-aware contrastive learning for cloth-changing person re-identification.
Explore all projects · Read my publications · View my CV
Research record
- Peer-reviewed work: WACV 2024 and RIVF 2025.
- Current research: Dynamic and large-scale graph community detection at the Institute of Mathematics, Vietnam Academy of Science and Technology, and graph-based spam detection with HUST.
- Applied impact: Graph-based COVID-19 risk assessment developed while working within Vietnam’s National Steering Team during the fourth wave.
- Recognition: Champion, MLOps Marathon 2023; Third Prize, Vietnamese Mathematical Olympiad 2017.
Contact
Email · GitHub · Google Scholar · LinkedIn · ORCID