I'm an AI/ML PhD student in the AIMS CDT. Most of my PhD work focuses on decoding speech from the brain (i.e. brain-to-text) using deep learning on non-invasive neuroimaging data in the Parker Jones Neural Processing Lab (PNPL). In the not so distant future, we hope this technology could assist paralysed patients to communicate via scalable brain-computer interfaces (BCIs). In the longer term, it could facilitate a plethora of applications that we are yet to understand.
I've also worked on open-ended learning through reference-free language model RL at Meta, and long-context reasoning at Google DeepMind. I remain interested in open problems in AI development, including continual and open-ended learning, creativity, and how we can imbue models with higher levels of agency.
