I'm interested in scientific machine learning, computational physics, NLP, and applied AI systems. Some repositories are learning projects or prototypes; others are ongoing research projects. Together they reflect the areas I am actively working in: simulation, model evaluation, data pipelines, and building tools that connect machine learning with real scientific and technical questions.
- Scientific machine learning and physics-informed modelling
- Unsupervised learning for simulated physics systems
- Building practical AI tools with clear evaluation and limitations
-
universality-discovery
Exploratory research notebook on whether unsupervised methods recover universality-class structure from simulated physics data. -
ml-universality-classification
Anomaly-detection experiments for simulated surface-growth universality classes. -
evaluation-cue-isolation-ai-agents Controlled experiments on evaluation-like prompt cues in AI-agent behaviour.
-
academic-literature-search
Applied AI search tool combining academic APIs with summarisation and retrieval.
I'm looking to build stronger engineering habits around research code: reproducible experiments, tests, clear documentation, honest limitations, and projects that can be understood by someone other than me six months later.