About

I am a research engineer working on large language model pretraining and optimization. Currently at OpenEuroLLM and the ELLIS Institute, my work focuses on training mechanics, scaling laws, and distributed training setups using Megatron-LM.

Before moving into full-time research, I spent two years as an ML engineer at SoftCo developing Bayesian inference and time-series models for fintech applications. I hold an MSc in Machine Learning from the University of Tübingen and a background in mathematics from Trinity College Dublin.

Education

  • MSc in Machine Learning (Grade 1.3) — University of Tübingen
    Thesis on adaptive and geometry-aware optimizers, supervised by Dr. Antonio Orvieto.
  • BA in Mathematics (First Class Honors, Gold Medal) — Trinity College Dublin
    Thesis on category theoretical applications in algebraic topology.