CV

The full version can be downloaded above.

Basics

Name Tom A. Lamb
Label PhD Student in Machine Learning, University of Oxford
Email thomas.lamb@eng.ox.ac.uk
Url https://tomalamb.github.io
Summary My research focuses on AI safety, robustness, and reliable machine learning. This includes work on certifying LLM judges and the relative calibration of teacher–student models, calibrating semantic confidence scores, and developing adaptive inference-time steering methods, alongside contributions to hallucination detection and universal in-context approximation for recurrent models.

Work

  • 2026.05 - 2026.10

    Paris, France

    Research Scientist Intern
    Apple Machine Learning Research
    Working with Sinead Williamson and Michael Kirchhof on post-training for budget-aware agents.
    • Internship culminated in the paper ‘Rewards for Budget-Aware Multi-Turn Agents That Adapt to Arbitrary Test-Time Constraints’.
  • 2022.07 - 2023.01

    Oxford, UK

    Machine Learning Research Intern
    DeepMind / Torr Vision Group, University of Oxford
    Designed a framework for faithful knowledge distillation. Derived LP-based bounds on teacher–student confidence disagreement enabling formal analysis. Developed methods yielding more calibrated student models; resulted in our paper Faithful Knowledge Distillation.
  • 2019.07 - 2019.09

    Durham, UK

    Summer Research Student
    University of Durham
    Proved new identities relating polylogarithmic integrals to multiple zeta values with Prof. Herbert Gangl.

Education

  • 2023.10 - Present

    Oxford, UK

    PhD (DPhil)
    University of Oxford
    Machine Learning, Department of Engineering Science
    • Supervisors: Tim G. J. Rudner (University of Toronto & Vector Institute) and Philip H. S. Torr (University of Oxford)
    • Research on AI safety and reliable ML, developing methods for LLM uncertainty, calibration, and generalisation.
  • 2022.09 - 2023.08

    Edinburgh, UK

    MSc
    University of Edinburgh
    Artificial Intelligence (Distinction)
    • MSc Artificial Intelligence Class Prize (ranked 1st overall)
    • Dissertation: Self-Supervised Learning of Tractable Generative Models
  • 2016.10 - 2020.07

    Durham, UK

    MMath
    University of Durham
    Mathematics (First-Class Honours)
    • Percy Heywood Prize; John Crowther Prize
    • Dissertation: An Introduction to Modular Forms and the Eichler–Shimura Isomorphism

Awards

  • 2023
    MSc Artificial Intelligence Class Prize
    University of Edinburgh
    Awarded to the student who attained the highest overall mark in the MSc Artificial Intelligence course.
  • 2020
    Percy Heywood Prize
    Department of Mathematical Sciences, University of Durham
    Awarded to a student graduating with an MMath whose performance is outstanding in the final year.
  • 2019
    John Crowther Prize
    University College, University of Durham
    Awarded for performance in third-year Mathematics examinations.