Computer Science student at Queen's University Belfast, interested in machine learning, research, and systems.

• Amadeus Counterpoint — investigating compositional behavioral modeling under unseen dyadic interaction, using chess as a controlled testbed.
• ChoiceBench — first-author research on accuracy and order sensitivity in multiple-choice LLM evaluation, currently available as an arXiv preprint and submitted through ACL Rolling Review.
• Building neural models and supporting infrastructure from first principles, with a current focus on chess, transformers, and representation learning.

• ChoiceBench
  A reproducible framework for evaluating positional and option-order sensitivity in multiple-choice LLM evaluation across models, datasets, and mitigation strategies.

• Amadeus Counterpoint
  A neural chess research project studying whether independently learned behavioral representations of two players can compose to reproduce their unseen interactions.

• Tokenizer & Language Model from Scratch
  Byte-level BPE tokenization and decoder-only transformer work focused on understanding the full language-modeling stack from first principles.

• Susume
  An anime recommendation system exploring collaborative filtering and matrix-factorization approaches.

Machine learning · Behavioral modeling · Representation learning · LLM evaluation · Neural chess · Robustness

Python · PyTorch · Java · NumPy · Linux · Bash · Git · LaTeX · C++