Robert is a PhD student in the EPSRC DRIVE-Health CDT, co-supervised by Dr Angus Roberts (King’s College London) and Dr Julia Ive (University College London). His research focuses on developing models that predict clinical trajectories and outcomes from Electronic Health Records, with an emphasis on creating realistic, stratified synthetic data to improve model robustness and generalisability. His work explores a range of modern generative and sequence modelling approaches, including autoregressive Transformers, large language models, state-space architectures, and knowledge graph representations.

Before joining the CDT, Robert worked at Amazon Web Services, contributing to the development of the company’s hypervisor and virtualization infrastructure. He previously completed an MSc in Artificial Intelligence with Machine Learning at Queen Mary University of London, where he proposed a novel method for Neural Architecture Search that integrates continuous pruning strategies with progressive network growth. He also holds a BSc in Computer Science from Nordakademie, where for his thesis he partnered with Continental AG in Germany to design a real-time, low-power FPGA-accelerated autonomous driving path planner.

Research interests

Deep Learning, Classical ML, Neural Architecture Search, LLMs, Gradient optimizers, Real-world application of AI/ML, Algorithm design and optimization

Hobbies and Interests

Travel, Dogs, Food, Transport&EVs, Energy, Trading, Politics, Video Games, Japan, Poland, More research