Research
Research into reasoning over structured systems.
My research explores how AI systems can reason more effectively over structured environments. I am particularly interested in knowledge graphs, graph representations, LLM reasoning, and the role of machine-understandable world models in making AI systems more capable and useful.
Current Focus
Improving reasoning over structured systems using graph-based representations.
My PhD focuses on improving reasoning over structured systems using graph-based representations. An early phase of the work explored LLMs for graph reasoning, including representation alignment, benchmarking, supervised fine-tuning, and the limitations of current approaches. This has since evolved towards graph-based world models and structured reasoning under uncertainty.
Research Themes
- Multi-agent reasoning over structured systems
- Graph representations and world models
- Representation alignment
- Structured reasoning under uncertainty
- Machine-understandable models of complex environments