Email: tharis at bu.edu

I am a third-year PhD student at Boston University, jointly advised by ***Krzysztof Onak*** and Venkatesh Saligrama. I did my undergrad at Dartmouth College advised by Amit Chakrabarti, and spent two years in Seattle as a software engineer at the Microsoft Windows Base Kernel team.

In 2025, I completed an internship at the National Institute of Informatics in Tokyo, hosted by Yuichi Yoshida.

I am currently a student researcher at Google Research, kindly hosted by **Maryam Karimzadehgan.**

My research mainly focuses on developing efficient and robust algorithms for Foundation Models. I also am often interested on problems from different areas of theoretical computer science, including sublinear algorithms, learning theory and complexity.

My CV

🎓 Google Scholar

Publications

Foundation Models

1. Noise Stability of Transformer Models: TH, Zhang, Yoshida (ICLR 2026)
2. Compression Barriers for Autoregressive Transformers: TH, Onak (COLT 2025)
3. $k$NN Attention Demystified: A Theoretical Exploration for Scalable Transformers: TH (ICLR 2025)
   • Poster Page

Sublinear Algorithms

1. Fast-MWEM: Private Data Release in Sublinear Time: Manuscript
2. Estimating Hitting Times Locally At Scale: TH, Spaeh, Dragazis, Tsourakakis (NeurIPS 2025)
3. Efficient Algorithms for Adversarially Robust Approximate Nearest Neighbor Search: Andoni, TH, Kelman, Onak
   1. NeurIPS 2025 Workshop: Reliable ML from Unreliable Data
   2. WoLA 2026 Poster
4. Counting Simplices in Hypergraph Streams: Chakrabarti, TH (ESA 2022)

Others

1. 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)

Talks

1. On Language Generation in the Limit [Oral Exam, BU]
2. On $k$-NN Methods for Fast Attention Computation [Boston University AI Seminar]
3. On Adversarially Robust Algorithms for Searching [Boston University Theory Seminar]