Tom A. Lamb

Tom A. Lamb is a DPhil (PhD) candidate in machine learning at the University of Oxford and a Research Scientist Intern at Apple Machine Learning Research (MLR) in Paris. His research interests span (i) AI safety and reliable machine learning, (ii) uncertainty quantification and calibration, (iii) generalisation of LLMs, and (iv) trustworthy agentic systems. His work includes semantic calibration and uncertainty estimation for LLMS, adaptive inference-time steering, robustness and bias in vision-language models, and PAC-Bayes generalisation guarantees for LLM safety monitors. He holds an MSc in Artificial Intelligence from the University of Edinburgh and an MMath in Mathematics from the University of Durham.

Personal Website

https://tomalamb.github.io/

LinkedIn

LinkedIn

Projects

No grants recorded.