Currently

• I’m a PhD candidate at UCL, advised by Gabriel Brostow and Jack Parker-Holder, as part of the UCL PRISM Lab, and also as part of UCL DARK.
• My research is at the intersection of computer vision, human computer interaction, and reinforcement learning. Lately, I’m trying to train models which generate novel environments.
• I’m also interested in how we can learn abstractions which can improve reasoning. I previously co-authored some thoughts on that here, but am now particularly focused on visual representation learning for reasoning.

Previously

• I co-founded OneContext , the RAG (retrieval augmented generation) pipeline company.
• I hold an MA (Oxon.) from Balliol College, Oxford , where I was a Nind Scholar, and an MS from University College London , where I graduated Dean’s List (top 5% of faculty). At UCL, I was fortunate enough to be advised by Marc Deisenroth .
• During my MSc, I was a researcher in The UCL Sustainability and Machine Learning Group and The Mercedes AMG F1 Team . My work for Mercedes is proprietary, but involved using Gaussian Processes to predict optimal times to release cars from pit-boxes during qualifying sessions.
• I’ve spent some time developing and trading systematic strategies in variance swaps, volatility swaps, and VIX options, at Capula (a $27bn quant hedge fund), and at J.P. Morgan Chase.
• I’ve been a non-executive director and AI advisor at a few interesting companies, across a range of industries including RevOps, Logistics, and PropTech. I mainly come in to help teams train and deploy AI models in safe and interpretable ways. Due to prior commitments, I am unlikely to take on any long-term (commercial) projects at the moment, but you can book time to chat with me here.

Teaching

• TA for COMP0137 - Machine Vision, and for COMP0258 - OpenEndedness & General Intelligence, Postgraduate courses offered in the following UCL Master’s Programmes: Machine Learning, Data Science & Machine Learning, Computational Statistics and Machine Learning, Computer Graphics Vision & Imaging.
• Helping run a multimodal intelligence tutorial in Summer 2026 at The Eastern European Machine Learning Summer School , with Skanda Koppula , in Montenegro.

If we share similar research interests, please reach out .