Hyemin Helen혜민 Bang

Hi! I'm Hyemin (Helen), a PhD student in EECS at MIT, advised by Mitchell Gordon. I work on what I call human-grounded interpretability: studying what's going on inside a model not as an end in itself, but so that people can trust, oversee, and collaborate with AI systems.

I approach this from two angles. One looks inward, using causal interpretability methods to uncover what and how models actually compute, tracing the internal processes behind their reasoning. The other turns outward, building interactions that put those internals within human reach, so that people's judgment, and their disagreement, can meaningfully steer how a model behaves.

Before my PhD, I earned my bachelor's and master's in Computer Science at MIT, working with Arvind Satyanarayan in the Visualization Group on interpreting and aligning model behavior with human reasoning. Before that, I spent two years as a systems developer at InterSystems, building ML integrations and data pipelines for their database systems.

Outside of research, I'm a Communication Fellow at MIT EECS, helping fellow students sharpen their writing and talks. I love mentoring, including undergraduates in data science and AI through Break Through Tech AI, and I've had the joy of TA'ing courses across AI and HCI at MIT.

When I'm not at my desk, you'll find me on a morning run, skateboarding to campus, tending a small garden, or hunting for vinyl. And wherever I am, there's a good chance my dog Chu is nearby :)