Tiannuo Yang

Tiannuo Yang (杨天诺)

Github / Google Scholar / CV / X / tiannuoy[at]usc[dot]edu

Hi there! I am an incoming PhD student at University of Southern California, advised by Willie Neiswanger. I work on System4AI & AI4System. I like to build low-cost, high-efficiency machine learning systems that serves various scenarios (e.g., LLM and RAG). Some of them have been deployed in industry.

I'm always open to discussions, collaborations, and speaking opportunities on AI Systems! Please contact me directly if interested!

News

[June 2025] Gave a talk on "Towards Efficient LLM Search Agents" at ByteDance, CUHK, and Nankai University. Thanks for the invitation!
[May 2025] Excited to share SearchAgent-X, a LLM-based search agent system aiming for EFFICIENCY. If you are post-training/deploying your own search agents, check it out!
[Mar. 2025] Thrilled to be pursuing my PhD at USC on AI Systems!
[Jan. 2025] One paper on Performance Tuning accepted by WWW 2025 as an Oral Presentation.
[Oct. 2024] Served as a reviewer for WWW 2025.
[March 2024] One paper on Performance Tuning accepted by ICDE 2024. See you in Utrecht!
[Sep. 2023] Got the first citation (by Cai et al., 2023) in my career!
[Sep. 2023] Released a collaborative Survey on Green Computing.
[June 2023] Started academic internship at Ant Group (Beijing).
[June 2023] One paper on Performance Tuning accepted by ICPP 2023.
[June 2023] Undergraduate thesis on Operations Research accepted by SEPS.
[Dec. 2022] Won the 3rd prize in Massive Storage Competition (organizer: Huawei).

Publications

Supercharging AI Systems

Demystifying and Enhancing the Efficiency of Large Language Model Based Search Agents
Tiannuo Yang, Zebin Yao, Bowen Jin, Lixiao Cui, Yusen Li, Gang Wang, Xiaoguang Liu
Arxiv Preprint, 2025
PDF | Code | BibTex

SCOOT: SLO-Oriented Performance Tuning for LLM Inference Engines
Ke Cheng, Zhi Wang, Wen Hu, Tiannuo Yang, Jianguo Li, Sheng Zhang
The Web Conference (WWW), 2025
PDF | Code | BibTex
(Oral Presentation)

VDTuner: Automated Performance Tuning for Vector Data Management Systems
Tiannuo Yang, Wen Hu, Wangqi Peng, Yusen Li, Jianguo Li, Xiaoguang Liu, Gang Wang
International Conference on Data Engineering (ICDE), 2024
PDF | Code | BibTex
(Deployed on Ant Group's CodeFuse)

CoTuner: A Hierarchical Learning Framework for Coordinately Optimizing Resource Partitioning and Parameter Tuning
Tiannuo Yang, Ruobing Chen, Yusen Li, Xiaoguang Liu, Gang Wang
International Conference on Parallel Processing (ICPP), 2023
PDF | BibTex

Survey

On the Opportunities of Green Computing: A Survey
You Zhou, …, Tiannuo Yang (Co-First Author), … and Xiaodong Zeng
ArXiv Preprint, 2023
PDF | BibTex

Talks

Invited talk on "Towards Efficient LLM Search Agents"

by Di Wu, at ByteDance, June 6th 2025;
at MLSys Reading Group, Nankai University, June 5th 2025;
by Haosen Shi, at CUHK, June 4th 2025

Presented work on VDTuner at ICDE 2024, the Neitherlands, May 2024
Presented work on CoTuner at ICPP 2023, Online, August 2023

Education

[from Aug. 2025] PhD Student of Computer Science, University of Southern California
[Aug. 2022 - June 2025] Master of Computer Science, Nankai University (exempted from the entrance exam)
[Aug. 2018 - June 2022] Bachelor of Information Management and Information System, University of Science and Technology Beijing (Major GPA: 3.97/4.0)
[Sep. 2019 - Jan. 2020] Exchange Program of Information Management, Southern Taiwan University of Science and Technology (GPA: 4.3/4.3)

Internship

[June 2023 - Jan. 2024] Academic Internship, Ant Group, Beijing, China

Honors & Awards

PhD Student Fellowship, University of Southern California, 2025
Excellent Graduate (8 out of 157), Nankai University, 2025
1st-Class Gongneng Scholarship, Nankai University, 2023, 2024
National 3rd Prize, Massive Storage Competition, Huawei, 2022
Provincial (Tianjin) Gold Prize, 'Internet+' Innovation and Entrepreneurship Competition, 2021
National 2nd Prize (1493/41826 = 3.6%), Contemporary Undergraduate Mathematical Contest in Modeling, 2020

Besides Research

Take a look at my Life Outside of Research.