Project lead on a UKRI government grant awarding up to £300,000 in GPU hours on the Isambard-AI supercomputer.
I am a PhD researcher at the University of Cambridge, work on autonomous vehicles at Minimal, writer read by 58,000+ people, and occasional ultramarathon runner.
I am a PhD researcher at the University of Cambridge, funded by the Medical Research Council industrial Collaborative Award in Science and Engineering scholarship.
I work on vision-language models. I am building end-to-end models that interpret social interaction, and deploying them to run on wearable glasses at low latency (100 ms). I conduct mechanistic interpretability on how these models represent a social scene. We collaborate closely with AI researchers at Stanford and Harvard.
Publications are available on Google Scholar.
How accurate are LLMs on multimodal theory of mind tasks
Natural language processing for analysis of electronic health records
I work closely with Emteq Labs. They make non-invasive human-computer interfaces.
I am working on a vision-language-action model for an autonomous delivery robot. I am building the pipelines for data collection and training, fine-tuning the model for deployment, collecting synthetic data, and developing the simulation environment.
I write about AI and neuroscience for 58,000+ people on Bluesky and Substack. I run workshops for teachers on AI and neuroscience in schools across the UK, and I recently delivered them to charities in Kolkata.
I was shortlisted for the Medical Research Council's Max Perutz Science Communication Award.
I host Bridging the Gap, a podcast of interviews with scientists.
I am currently raising money for sarcoma.
101km, with 3,000m of elevation, Peak District ultramarathon with Sacha.
800km in 8 days for Alzheimer's. We raised £7,000. Follow the journey here.
I will be running the London Marathon and the Cotswold ultramarathon in 2027.
I'm also converting a classic VW T3.