Matt McManus

Matt McManus

MIT CS & Math

Research-oriented ML engineer focused on structure-aware learning and reliable decision systems.

I focus on leakage-safe evaluation, probability calibration, and mechanistic probes for LLM agents in macro settings.



Research Interests: Probability calibration, mechanistic interpretability, scientific machine learning, leakage-safe evaluation.


Outside of work, I love working out, playing basketball, and keeping up with the latest AI/ML research.