I'm a Machine Learning Engineer and Researcher based in London.
Following an MSc in Data Science and Machine Learning at @UCL, I joined @Pontikos Lab as a Research Assistant, where I focus on Computer Vision and Synthetic Data Generation. Previously, I spent three years in Data Science within Wealth Management and interned at @Imagination Technologies
I am fascinated by Computer Vision, Generative Models, and Self-Supervised Learning. I am particularly interested in Synthetic Data and the emerging field of World Models for real-world applications, ranging from autonomous systems to Embodied AI.
Outside of ML research, I enjoy all things Basketball, MotoGP, video games, and hiking. These passions often serve as the testing grounds for my personal projects.
- MotoReID
An end-to-end computer vision pipeline utilizing YOLOv8 and DINOv3 (Vision Transformer) embeddings for high-speed sports re-identification. The system solves for persistent identity tracking across extreme occlusions and motion blur. - SiT FAF Generation
A generative modeling framework for synthetic medical image synthesis, inversion, and semantic editing using Scalable Interpolant Transformers (SiT). It enables conditional generation based on genetic mutations laterality, and patient age. - Semantic Context Tokens
Developed a coarse-to-fine tokenization pipeline integrating semantic tokens with subword units to enhance LLM narrative coherence. Inspired by Meta’s Large Concept Model, this approach yielded significant gains on the TinyStories benchmark. - Steven Medical Copilot
A medical voice assistant prototyped for the ElevenLabs x a16z AI Agent Hackathon. It leverages NLP and conversational AI to automate clinical documentation and referral letter composition.
Other notable projects include Contra-CTGAN and WeakTR Refinery.