Isabel MacGinnitie

My interests lie in doing thoughtful AI alignment research, both technical (building AI models that do what we tell them) and normative (building models that do what is morally right (and often implicitly, choosing what we think is morally right)). I’m currently a research fellow at Athena and SPAR. I am also applying to CS PhD programs for Fall 2026, supported by the NSF CSGrad4US fellowship.

My work is guided by these overarching questions:

• How can we use our theories and techniques from social sciences to do better alignment research?
• On the technical side, how can we use our understanding of emergent model behavior to improve our training and monitoring, so we can build more capable and robust models?
• On the normative side, how can we use moral philosophy and applied ethics to improve value conflict resolution and moral reasoning, so we can design robust AI systems able to navigate ethical ambiguity and complex value conflicts in novel domains?

Previously, I was make getting more people into homes by accelerating mortgage assumptions as first engineering hire at RetroRate, and improving access to early cancer detection as a software engineer at Color. I graduated Harvey Mudd College in 2023 with a BS in computer science and math with high distinction.