I study student modeling and simulation, with a focus on how heterogeneous educational data — dialogue, behavioral logs, questionnaires, and classroom interactions — can be integrated to represent learner characteristics and developmental trajectories.
LLM-based student agents & agent-based modeling
Individual-level behavior simulation
Learning processes reflected in dialogue and interaction data
Interpretability and theoretical alignment in AI systems
Epistemic boundaries of generative models in education
Rather than treating AI as a black-box predictor, I focus on what can — and cannot — be meaningfully simulated at the level of individual learners. My work explores the tension between observable behavior, latent constructs, and language-based representations of cognition. Reach my work through ArXiv!
Alongside research, I design and implement full-stack educational systems, including asynchronous task orchestration, progress tracking, and real-time feedback pipelines — examining how technical architecture and theoretical assumptions co-evolve in AI-supported learning environments. Checkout Feedforward System and several tools on PyPI!
When I’m not working on models or systems, I enjoy:
🏀 Basketball, badminton or swimming
📱 Clash Royale
💻 The Binding of Isaac, Mahjong Soul
Reach me or my team through personal email, ffs_tech or feal_tech.