Pragya Singh
DPhil student, Department of Statistics, University of Oxford
Email · Google Scholar · GitHub · LinkedIn
Hi, I am Pragya, a first-year DPhil student in Statistics at the University of Oxford, supervised by Prof. Yee Whye Teh and Prof. Volker Deringer. I am part of the Intelligent Earth CDT, which I joined in 2025 for a year of doctoral training before starting my DPhil research in October 2026.
My current work focuses on generative and sampling methods for amorphous materials. Before this, I worked on geometric deep learning. I developed graph neural networks for brain graph super-resolution with Prof. Islem Rekik, and invariant architectures for the Calabi–Yau dataset in string theory with Dr. Daniel Platt. I also explored topology-aware diffusion models with Prof. Bastian Rieck.
Before my MSc, I spent three years as a quantitative researcher at Goldman Sachs.
Publications
-
Rethinking Graph Super-resolution: Dual Frameworks for Topological Fidelity
P. Singh, I. Rekik
arXiv preprint, 2025 -
Strongly Topology-preserving GNNs for Brain Graph Super-resolution
P. Singh, I. Rekik
PRIME Workshop, MICCAI 2024 -
Group-invariant Machine Learning on the Kreuzer–Skarke Dataset
C. Ewert, S. Magruder, V. Maiboroda, Y. Shen, P. Singh, D. Platt
Physics Letters B, 2024
Education
-
2025 – present
DPhil in Statistics, University of Oxford
Intelligent Earth CDT · Supervisors: Prof. Yee Whye Teh & Prof. Volker Deringer -
2023 – 2024
MSc in Artificial Intelligence, Imperial College London
Distinction · Thesis with Prof. Islem Rekik, awarded best MSc thesis in the cohort -
2016 – 2020
B.Tech in Engineering Physics, IIT Roorkee
Minor in Mathematics
Experience
- 2025 – 2026
-
2020 – 2023
Quantitative Researcher, Goldman Sachs, Bengaluru
Funding analytics and optimisation for firm-wide liquidity planning