I am a Ph.D. candidate in the Department of Computer Science at the University of California, Santa Barbara. I am co-advised by Professor Divyakant Agrawal and Professor Amr El Abbadi. I'm interested in developing new algorithms and systems to manage big data efficiently. Currently, I am exploring the opportunities in applying large language model in data systems (such as translating natural language to SQL).
Besides research, I enjoy trying new food, watching movies, and playing tennis and badminton. I'm also a trumpet player in which I played trumpet solo in Danzón No. 2 at the Disney Hall.

I have previously interned at Snowflake and Microsoft Gray System Lab. Contributions to FoundationDB (while interning at Snowflake) can be found at Github.
Thanks for the Microsoft Research PhD Fellowship from Microsoft Research and gifts from Snowflake Inc.

Interested in collaboration? Drop me an email and let's chat!

Publications and Ongoing Projects

A Detailed Analysis of the SpaceSaving± Family of Algorithms with Bounded Deletions

Fuheng Zhao, Divyakant Agrawal, Amr El Abbadi, Claire Mathieu, Ahmed Metwally, Michel de Rougemont

preprint

NL2SQL is a solved problem... Not!

Work done while interning at Microsoft Gray System Lab

CIDR 2024

Panakos: Chasing the Tails for Multidimensional Data Streams

Fuheng Zhao, Punnal Ismail Khan, Divyakant Agrawal, Amr El Abbad, Arpit Gupta, Zaoxing Liu

VLDB 2023

Differentially Private Linear Sketches: Efficient Implementations and Applications

Fuheng Zhao*, Dan Qiao*, Rachel Redberg, Divyakant Agrawal, Amr El Abbadi, Yu-Xiang Wang

NeurIPS 2022

SpaceSaving±: An Optimal Algorithm for Frequency Estimation and Frequent Items in the Bounded Deletion Model

Fuheng Zhao, Divyakant Agrawal, Amr El Abbadi, Ahmed Metwally

VLDB 2022

KLL±: Approximate Quantile Sketches over Dynamic Datasets

Fuheng Zhao, Sujaya Maiyya, Ryan Wiener, Divyakant Agrawal, Amr El Abbadi

VLDB 2021

Autumn: A Scalable Read Optimized LSM-tree based Key-Value Stores with Fast Point and Range Read Speed

Looked into the trade-off between write and read. Proposed a new ganerning compaction policy to enhance read performance .

preprint