About Me

Hi! My name is Sai Avula. I'm interesting in building safe, reliable, and verifiable systems for AI agents.

After receiving my BS in Computer Science from RPI - I worked at The MITRE Corporation as a Data Scientist where I leveraged advanced statistical analysis and build machine learning tools to tackle complex public policy problems such as healthcare, data privacy, welfare, and supply chain logistics.

As a graduate student at The University of Michigan, I co-authored a <a href = https://arxiv.org/abs/2511.02109>NeurIPS 2025 Spotlight Paper</a> developing a framework to evaluate agent alignment with human values for decision-making.

Recently, I developed Distillation Guard, a monitoring service coupled with a codebase knowledge agent that uses a classifer trained on distillation attacks to modify agent behavior if unwanted capability extraction occurs. This project won 3rd place at LightningAI and Validia's Personalized AI Agent Security Hackathon in NYC!

Overall, my work connects technical AI research with real-world impact, such as training climate forecasting models with informative uncertainty quantification outputs, developing privacy-preserving tools, and designing accessible AI tools and policies. My current research interests include mulit-agent orchestration, AI governance, scalable network analysis, and open source data provenance protocols.

In my spare time, I love traveling and sampling new cuisines, playing pickup basketball, producing music, and backpacking in the mountains!

I am currently looking for full time for Data Scientist and AI Engineer roles. If you are looking to connect with me an opportunity or if you want to chat about any of my interests please contact me through email or LinkedIn.

Below are some pics from my favorite adventures!