Amirhossein Ghaffari

I am a doctoral researcher in Computer Science and Engineering at the Future Computing Group (FCG), Centre for Applied Computing, Faculty of Information Technology and Electrical Engineering (ITEE), University of Oulu, Finland, supervised by Adj. Prof. Ekaterina Gilman and Adj. Prof. Susanna Pirttikangas. I develop graph-based machine learning methods that turn heterogeneous real-world data into reliable predictions and decision support for cities and mobility systems.

My PhD focuses on spatio-temporal graph neural networks (STGNNs) for urban prediction, combining ideas from machine learning, network science, and urban analytics. A central theme of my work is how raw spatio-temporal data should be mapped into graphs, and how model design can make graph learning more accurate and robust. This line of work includes road-aware graph construction (STRAM, Neurocomputing), the open-source STM-Graph framework (CIKM 2025), and mixture-of-experts GNNs for traffic forecasting (IEEE MDM 2026), with further work on graph transformers and spatio-temporal model design in progress.

From September to November 2026, I am a visiting doctoral researcher with the Machine Learning for Smart Mobility (MLSM) group at the Technical University of Denmark (DTU), hosted by Prof. Francisco Camara Pereira.

Beyond GNNs, I work on LLMs and agentic AI, including adapting LLMs to online communities (RedditPersona, CIKM 2026), LLM-powered travel planning from mobility patterns (IEEE BigData 2025), and agentic systems built on knowledge graphs. I have also contributed to research on federated and privacy-preserving learning, smart city optimization, and anomaly detection in 5G networks.

Before my PhD, I spent over three years as a machine learning engineer at Fard Iran Co., building production AI for road traffic: automatic number plate recognition, weigh-in-motion (WIM) vehicle weighing, intelligent surveillance, and driver-behavior analysis. I helped deploy and maintain these systems at more than 70 sites. In parallel, I developed deep learning models for air pollution prediction at the Institute of Water and Energy, Sharif University of Technology. I hold an M.Sc. from Sharif University of Technology and a B.Sc. from Amirkabir University of Technology.

I enjoy the full path from a research idea to a system people actually rely on. I am always happy to discuss research, collaboration, or new opportunities. Feel free to reach out via email or LinkedIn.