Samuel Wu (he/they)

About me
I recently finished my master's in High Performance Computing (HPC) with Data Science at the University of Edinburgh. My master's dissertation focused on replicating and extending DeepPCR, an algorithm from a recent NeurIPS paper that applies parallel tridiagonal solvers to neural networks. I hold a bachelor's degree with honours in Computer Science from Lancaster University, with a full exchange year at the Australian National University.
I am interested in HPC in general and, in particular, leveraging accelerators for machine learning (ML) applications. Currently, I am focused on numerical methods for ML and generative models (normalising flows, flow-matching).
Outside of my academic pursuits, I enjoy reading and watching films. I am also actively involved in the community, having given talks at Yorkshire DevOps and BSides Leeds.
Updates
Aug 2025Article: The Unsubmitted 2nd-Place LLM Inference Solution
Aug 2025I have completed my MSc dissertation! [View abstract]