Faculty Market Portfolio
Statements from my faculty applications.
About Me
I am an Assistant Professor in Boston University’s faculty of Computing & Data Sciences interested in language model training mechanics: how models learn to encode linguistic patterns or other structure and how we can encode useful inductive biases into the training process. I also work broadly on interpretability and robust generalization. Previously, I earned a PhD from the University of Edinburgh on Training Dynamics of Neural Language Models, was a postdoc at NYU, and was a Kempner Research Fellow at Harvard. My work has been profiled in Quanta and The Register. I was named a Rising Star by MIT EECS and awarded the Google Europe Scholarship for Students with Disabilities, the latter in recognition of my reliance on code dictation. Outside of research, I play roller derby under the name Gaussian Retribution and perform standup comedy.
I am recruiting PhD students to begin in 2027 at Boston University. Do not email me before reading my contact notes.
Interests
- Language modeling
- Interpretability
- Training Dynamics
- Generalization
- AI for Scientific Understanding
Education
PhD in Informatics
University of Edinburgh
MEng in Computer Science
Johns Hopkins University
BSc Artificial Intelligence
Carnegie Mellon University
My Research
I want to completely and comprehensively understand language model training. This objective combines linguistics, optimization, learning dynamics, science of deep learning, interpretability, and behavioral analysis. Recently, I have begun using similar approaches to study scientific discovery models and enhance broader scientific understanding.
My top three current research goals are:
- Leveraging training trajectories and variation between runs to identify what concepts are significant and distinct to a model. How do these concepts relate to each other?
- Predicting model behavior on edge cases by holistically understanding models on a computational and algorithmic level.
- Expanding human understanding of the world by studying how models learn to simulate it. To achieve this interdisciplinary objective, my current collaborations aim to deeply understand fish, stars, and weather as well as language.
My current publication list is available on my Google Scholar.
Recent Posts
The AI Researcher's Guide to a Non-Boring Bluesky Feed
How to migrate to bsky without a boring feed.
The Parable of the Prinia's Egg: An Allegory for AI Science
What counts as strong evidence for an explanation of model behavior?
Interpretability Creationism
Nothing in Deep Learning Makes Sense Except in the Light of SGD.
Potential Mentees
You do not need to email me to apply to my lab and it will offer most applicants no advantage to do so. One exception for applicants with existing work to share: If your work is highly relevant to me, feel free to reach out to discuss it. I am not currently hiring interns or looking for remote junior collaborators. I welcome any specific connections and questions about my research, though I may direct you towards my coauthors who did the real work related to your inquiry.
I welcome any messages from fellow disabled researchers looking to connect—I have direct personal experience in this arena.
Am I a good advisor for you?
If you like my papers, you may be a good fit! However, note that I am primarily interested in understanding systems, not creating them. Occasionally, my collaborators and I discover a new way to improve models; this is entirely by accident and should not be expected. If you are trying to build exciting new models or fix current ones, you may want to look for a different advisor.