Physics graduate currently studying a master's in artificial intelligence.
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
- NLP and computational linguistics research methods
- 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. -
PhysForge
Prototype web app for PDE discovery using physics-informed neural networks and sparse regression. -
ML_High_Jump_Performance_System
Ongoing biomechanics research collaboration with the current national champion in high jump & biomedical researcher at GMRI, exploring pose estimation, physics-informed modelling, and technique analysis for high jump. -
academic-literature-search
Applied AI search tool combining academic APIs with summarisation and retrieval. -
computational_linguistics_peer_review_analysis
Computational linguistic analysis of matched preprint and publication pairs. -
from-scratch-decision-tree-python
Small ID3-style decision tree implementation written without machine-learning libraries.
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.