PROJECTS

Over the past six years of researching in Higgs physics, I have learnt how the field and the questions it asks has evolved. This understanding has come from conversations with colleagues and mentors, as well as from the vast body of literature I have explored myself. With my training in Data Science, Machine Learning (ML) and Natural Language Processing (NLP), I began to wonder:

Can the collective knowledge of our field be systematically learned and visualized using modern machine-learning methods?

To explore this, I applied NLP techniques to hundreds of high-energy-physics abstracts and titles, building a pipeline that uncovers latent topic structures and quantifies their evolution through time. Methods such as BERTopic identify semantically meaningful clusters, while ARIMA forecasting models their temporal dynamics — revealing the rise and decline of subfields within particle physics.

The results provide a data-driven view of the sociology of modern physics, illustrating how theoretical and experimental communities have co-evolved over the decades and how research priorities continue to shift with discovery and technology.

This collection of projects was part of the University of Cambridge Data Science, Machine Learning & AI Career Accelerator. To keep the experience fair for future learners, the full project materials aren't publicly shared, but the outcomes and insights attached reflect my own work and learning journey.

1) Anomaly Detection

2) Customer Segmentation

3) Predictive Modelling

4) NLP Topic Modelling

5) Time-Series Forecasting.

Each project deepened my ability to turn complex data into insight — showing how strong analytical foundations and modern ML tools can make data both meaningful and actionable.