To appear in ACM International Conference on Management of Data (SIGMOD), 2027
Ruiying Ma
I am a second-year Ph.D. student at the UC Berkeley EPIC Data Lab, advised by Prof. Aditya Parameswaran. I received my B.Eng. in Computer Science from Tsinghua University (Yao Class) in 2025. Prior to this, I was a research intern at the Systems and Networking Research Group at Microsoft Research Asia.
News
- Sep 12, 2026 SHED (Robust Hierarchical Structures for Agentic Document Analysis) is accepted to SIGMOD 2027!
- Aug 20, 2026 DAB is accepted to the Main Conference of EMNLP 2026 (acceptance rate 15.4%)! The leaderboard has already received over 40 submissions.
- May 26, 2026 I gave an invited talk on DAB at ByteDance (San Jose).
- Mar 2, 2026 We release DAB, a new benchmark for data agents on realistic, multi-system data tasks. Contributions welcome!
Research
My research focuses on agentic data systems, especially how to help AI agents understand and process large-scale data in realistic, complex environments. In particular, I study two directions:
- Improving systems to better support AI agents: What is the right system interface and architecture, and how can we improve the system's performance?
- Designing agents to understand and process data more effectively: What is the right agent design, and how can we improve the agent's accuracy and efficiency in data processing?
Publications
* Co-first authors.
Empirical Methods in Natural Language Processing (EMNLP), Main Conference, 2026
International Conference on Learning Representations (ICLR), 2026
IEEE International Conference on Data Engineering (ICDE), 2025