I study internal representations in both biological and artificial neural systems: how they are formed, how they can be interpreted, and how they shape behavior.
I’m a Research Scientist with a background in Computational Neuroscience, currently focusing on Mechanistic Interpretability. My work sits at the intersection of neuroscience, machine learning, and complex systems: basically taking messy, complex data (rodent brain recordings or neural network activations) and turning it into insights that actually make sense.
- PhD research: Led full-cycle experiments and computational analyses to understand spatial representation in the mouse hippocampus (“the brain’s GPS”).
- Mechanistic interpretability: Completed a research project mentored by Stefan Heimersheim (at FARAI) on computation in superposition, resulting in a NeurIPS workshop paper and LessWrong article.
- Cross-domain representations: collaborated on a recent NeurIPS workshop paper applying Sparse Autoencoders to biological neurons, showing how interpretability tools can bridge AI and Neuroscience.
- Python: PyTorch, TransformerLens, NumPy, Pandas, Scikit-learn, Matplotlib, Seaborn
- Neural networks, FastICA, Bayesian Inference, SVM, Gradient Boosted Trees, random forests
- LLM experimentation, mechanistic interpretability
📫Find me on:
💻 Programming languages: