Monte Hoover

I am a PhD student in Computer Science at the University of Maryland, College Park, advised by Professors Tom Goldstein and Ramani Duraiswami. My research focuses on AI safety in language models and on long context reasoning and efficiency. In the safety space I work on guardrails for language models and agentic systems, with a focus on complex and nuanced scenarios where current guardrails fail. I also work on reasoning capabilities in sub-quadratic attention mechanisms with a focus on approximate nearest neighbors search and Taylor series attention kernels.

I have collaborated with researchers from Capital One Research and the Army Research Laboratory and my work is supported by a National Physical Science Consortium fellowship.

Before joining UMD, I worked at Microsoft where I led the Europe, Middle East, and Africa division of Azure Kubernetes Support Engineering with offices in Portual, Romania, Israel, and Jordan. Prior to that I worked on the TLC team in Microsoft Research. Before to my time at Microsoft I served as a Captain in the United States Army and was selected for a leadership development program at Goldman Sachs where I focused on improving notional pricing for structured products.

I hold a Bachelors of Science in Computer Science from the United States Military Academy. For more details, you can view my CV.

My latest project is on dynamic guardrails for language models:

Check it out at https://montehoover.github.io/DynaGuard/