About Me
I’m Greg, a machine learning researcher and football data scientist. I did my PhD at the University of Southampton and now work at Southampton FC. I build AI and machine learning models that help football clubs improve short and long-term performance. My work is grounded in peer-reviewed research and shaped by hands-on experience inside a professional football club, applying that research to real-world decisions in recruitment and on-pitch performance.
How I Got Here Link to heading
As a lifelong football fan studying computer science, I became fascinated by how the vast amount of data surrounding the sport could be used more effectively. That curiosity became the foundation of my PhD, where I set out to build AI and analytics that inform real-world decisions inside football clubs, not just predict who might win the league.
What started as academic research quickly became applied work. Throughout my PhD, I collaborated with Sentient Sports, keeping my research grounded in the practical realities of club operations. That experience led me to my current role at Southampton FC, where I translate research into real football decisions every day.
What I Work On Link to heading
Football clubs operate in a high-stakes environment with millions of pounds in player assets, constant pressure to balance performance with injury risk, and often incomplete information. These are exactly the kinds of messy problems AI should be helping with.
My research treats football teams as multi-agent systems: groups of players who need to coordinate effectively under uncertainty. I use techniques including reinforcement learning and graph neural networks to understand how teams collaborate and how individual contributions shape overall performance.
One project I’m particularly proud of is a team selection model that accounts for player injury risks. It demonstrated a potential ~13% reduction in first-team injuries while cutting the wages clubs inefficiently spend on injured players by ~11%, all without sacrificing on-pitch performance. For any club, that’s potentially millions saved and, more importantly, player careers protected.
I also work on making advanced analytics accessible to clubs without expensive tracking data. By developing methods to extract spatial insights from basic event data, I help level the playing field for clubs operating outside the top leagues.
How I Work Link to heading
AI works best when it augments human expertise rather than replacing it. Football people understand the game in ways no algorithm ever will. What AI can do is process and predict from huge volumes of data, surfacing patterns across thousands of matches and millions of events in seconds, work that would take a team of analysts months. Combining that scale of evidence with deep football knowledge is where the most valuable insights come from.
My experience at a professional club has reinforced this belief. The most impactful data solutions come from consistent collaboration between technical and football expertise. The combination is where the real edge lies.