CS undergrad · ML & open source
Curious about how models actually work. I contribute to open-source ML libraries, build projects to test ideas, and read a lot of things I don't fully understand yet.
Most of my learning happens outside class.
- Contributing to open-source scientific Python libraries — in active maintainer review
- Refactoring internals of an LLM tooling library
- Writing up a mechanistic interpretability project as a preprint
Things I keep coming back to: mechanistic interpretability, ML security, graph neural networks, scientific computing. Less "using models", more "understanding what's happening inside them."
Languages: Python · C++ (basics) · Go (basics)
ML: PyTorch · Scikit-learn · TransformerLens · SAELens · XGBoost
Ecosystem: NumPy · Pandas · FastAPI · Qdrant · Git · Jupyter
- Open-source contributions across scientific Python ecosystems — features and bug fixes, real review cycles
- Mechanistic interpretability experiments on a transformer model — feature isolation, causality verification
- Adversarial ML research — quantifying attack surfaces in neural network architectures
- Placed in top teams at a multi-team AI hackathon
- Picked up enough Go to ship a fix in a production CLI
