CV

Below is a brief outline of my CV. The full version can be found here: View Full CV

Research Experience

• Machine Learning Research Intern, Universtiy of Oxford (July 2022 - Sept. 2022).
  • Research intern funded by Google DeepMind and supervised by Francisco Girbal Eiras and Prof. Philip H. S. Torr.
  • Introduced a faithful‐imitation framework for teacher–student networks, formulated LP‐based upper/lower bounds on measures of the local confidence disagreement between teacher and student models, and designed a distillation method yielding verifiably more calibrated student models.
  • Lead to our paper ``Faithful Knowledge Distillation”.
• Summer Mathematics Research Student, University of Durham (July 2019 - Sept. 2019).
  • Collaborated with Prof. Herbert Gangl to conjecture and prove new identities relating polylogarithmic integrals over the unit square to linear combinations of multiple zeta values.

Education

• DPhil Machine Learning, University of Oxford (Oct. 2023 - Present)
  • Supervised by Prof. Philip H.S. Torr (University of Oxford) and Dr. Tim G.J. Rudner (NYU).
  • Published at top-tier ML conferences including ICML, NeurIPS and ICLR. These include papers on: the semantic calibration of and uncertainty quantification for LLMs; steering LLMs adaptively at inference time; biases in VLMs; and giving formal statements on universal in-context approximation of fully recurrent models including SSMs.
• MSc Artificial Intelligence, Univeristy of Edinburgh (Sept.2022 - Aug. 2023).
  • Courses taken included: Reinforcement Learning, Bayesian Statistics and MCMC Methods, Probabilistic Modeling and Reasoning, and Natural Language Understanding, Generation and Machine Translation. Audited Targeted Causal Learning, Methods of Causal Inference, and Algorithmic Game Theory.
  • MSc dissertation: `Self‐Supervised Learning of Tractable Generative Models’, supervised by Dr. Antonio Vergari. Applied conditional composite log‐likelihood estimation (CCLE) with novel patching schemes to train EiNet models. Demonstrated that training using CCLE with specific patching schemes shows promise for improved image inpainting performance.
  • Recieved the MSc Artificial Intelligence Class Prize.
• MMath Mathematics, University Of Durham (Oct. 2016 - July 2020).
  • Specialised in pure mathematics. Selection of courses taken: Algebraic Topology, Riemannian Geometry, Representation Theory, Partial Differential Equations, Algebraic Number Theory, Galois Theory, Statistics, Analysis in Many Variables, Algebra, and Numerical Analysis.
  • MMath dissertation: `An Introduction to Modular Forms and the Eichler–Shimura Isomorphism’, supervised by Prof. Herbert Gangl. Provided a comprehensive introduction to the theory of classical Modular Forms. This included detailed discussions on Hecke operators, Maeda’s conjecture, $L$-functions and was concluded by proving the Eichler–Shimura Isomorphism.
  • Recieved the John Crowther and Percy Heywood Prizes.