During my CDT summer placement at the ANGLE Lab, I researched Zero-Knowledge Machine Learning (ZKML): a cutting-edge field at the intersection of theoretical cryptography and applied artificial intelligence. Research Intern: Huy Quang Le Project supervisor: Rik Sarkar, Michele Ciampi PhD Tutor: Thomas Wong During my CDT summer placement at the ANGLE Lab, I researched Zero-Knowledge Machine Learning (ZKML): a cutting-edge field at the intersection of theoretical cryptography and applied artificial intelligence. My project addressed a major barrier to ZKML deployment: the highly unpredictable computational cost of compiling neural networks into cryptographic circuits. Utilising the university's Eddie HPC cluster, I empirically benchmarked thousands of models to develop a non-linear mathematical cost model and an empirical Pareto frontier, providing developers with actionable guidelines for optimising ZK circuits. Coming from a pure mathematics background at the University of Bristol, this placement catalysed immense technical and personal growth. Prior to this summer, my studies were focused primarily on abstract mathematical concepts, with little to no exposure to applied computer science or research literature. Over eight weeks, I learnt to navigate Linux terminals and High-Performance Computing environments, utilise version control, and train complex PyTorch models. I additionally completed university modules in Cryptography, Data Science and Machine Learning to bridge my knowledge gaps. Beyond the technical milestones, the vibrant research culture at the University of Edinburgh was profoundly inspiring. I had the privilege of presenting my project to established academics, iterating on my work based on their rigorous peer feedback. The welcoming culture of the School of Informatics (from attending visiting lectures to group outings and watching the World Cup together) showcased the collaborative, human side of academia. Ultimately, this placement fundamentally shifted my perspective by showing me the exciting intersection between abstract mathematical concepts and practical machine learning. Working alongside industry leaders and incredible fellow interns has ignited a clear ambition to pursue a PhD. In the immediate term, I am excited to continue collaborating with my supervisor on this project, with the aim of publishing our benchmarking framework at an upcoming conference. This article was published on Friday 30 October 2026
