About Me

I am a second-year PhD student at Boston University, advised by Krzysztof Onak. Before BU, I completed my undergrad at Dartmouth College, and spent two wonderful years as a software engineer at the Microsoft Windows Base Kernel team.

LinkedIn, Google Scholar

CV

Email: tharis at bu dot edu

Research Interests

My research lies at the intersection of theoretical computer science and artificial intelligence. I seek to understand how theoretically principled approaches can help us do Machine Learning sustainably, efficiently and securely. More specifically, I have been thinking about the following topics:

• Theoretical insights for LLMs and Transformers
• Learning Theory in the world of LLMs.
• Property testing.
• Efficient differentially private algorithms.
• Efficient algorithms on Markov Chains
• Adversarially robust randomized algorithms.

Publications

1. Compression Barriers for Autoregressive Transformers: TH, Krzysztof Onak (COLT 2025)
2. $k$NN Attention Demystified: A Theoretical Exploration for Scalable Transformers: TH (ICLR 2025)
   • Poster Page
3. Counting Simplices in Hypergraph Streams: Amit Chakrabarti, TH (ESA 2022)
4. Teaching American Sign Language in Mixed Reality: Shao, Sniffen, Blanchet, Hillis, Shi, H, Liu, Lamberton, Malzkuhn, Quandt, Mahoney, Kraemer, Zhou, Balcom (Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies)

Blog

I maintain a small blog of my notes and thoughts on various topics, technical or not, here.

Blog