Yuanli Wang (王远立)

I'm a Ph.D. candidate at Complex Analytics & Scalable Processing (CASP) Research Lab at Boston University, and I am working with Vasiliki (Vasia) Kalavri.

My research interests are in distributed systems and stream processing systems. I also work on Agentic AI systems and applications.

Email: yuanliw at bu dot edu  /  CV  /  Github  /  Google Scholar

profile photo

Publications
  • Agentic AI systems and applications
  • WHEN TO ACT, WHEN TO WAIT: Modeling Structural Trajectories for Intent Triggerability in Task-Oriented Dialogue
    Yaoyao Qian, Jindan Huang, Yuanli Wang, Simon Yu, Kyrie Zhixuan Zhou, Jiayuan Mao, Mingfu Liang, Hanhan Zhou
    Social Sim @ COLM 2025 (to appear) [Demo] [Code]
    VisualTreeSearch: Understanding Web Agent Test-time Scaling
    Danqing Zhang, Yaoyao Qian, Shiying He, Yuanli Wang, Jingyi Ni, Junyu Cao
    ECML-PKDD 2025 Demo Track (to appear) [Demo] [Playground]
  • Data stream processing systems
  • CAPSys: Contention-aware task placement for data stream processing
    Yuanli Wang, Lei Huang, Zikun Wang, Vasiliki Kalavri, Ibrahim Matta.
    EuroSys 2025 [code] [poster] [Tech report]
    The Non-Expert Tax: Quantifying the cost of auto-scaling in Cloud-based data stream analytics
    Yuanli Wang, Baiqing Lyu, Vasiliki Kalavri
    BiDEDE’22, co-located with SIGMOD’22 [slides] [poster]
    A New Benchmark Harness for Systematic and Robust Evaluation of Streaming State Stores
    Esmail Asyabi, Yuanli Wang, John Liagouris, Vasiliki Kalavri, Azer Bestavros
    EuroSys 2022 [code] [slides]
  • Machine Learning Systems
  • HACCS: Heterogeneity-Aware Clustered Client Selection for Accelerated Federated Learning
    Joel Wolfrath, Nikhil Sreekumar, Dhruv Kumar, Yuanli Wang, Abhishek Chandra
    IPDPS 2022 [code]
    Accelerated Training via Device Similarity in Federated Learning
    Yuanli Wang, Joel Wolfrath, Nikhil Sreekumar, Dhruv Kumar, Abhishek Chandra
    EdgeSys 2021, co-located with EuroSys 2021 [talk]
    Exploiting Data Heterogeneity for Performance and Reliability in Federated Learning
    Yuanli Wang, Dhruv Kumar, Abhishek Chandra
    Poster in the Fifth ACM/IEEE Symposium on Edge Computing. 2020
  • Distributed Systems
  • Fail-slow fault tolerance needs programming support
    Andrew Yoo, Yuanli Wang, Ritesh Sinha, Shuai Mu, Tianyin Xu
    HotOS-XVIII [talk]

    Work Experiences
  • Apple , AIML - Data Processing Platform Intern , 05/2023 - 08/2023
  • Integrate data lineage tracking support for Flink platform.

  • PingCAP , Database Engineer Intern , 05/2019 - 08/2019
  • Worked on AutoTiKV project from scratch: use machine learning to tune database under user-specific workloads.


    Invited Talks
  • Towards a cost-efficient and QoS-aware self-managed stream processing system, Meta, July.2022
  • CAPSys: Contention-aware task placement for data stream processing, Tufts University, Oct.2024

  • Professional Services
  • Reviewer for Journals: Future Generation Computer Systems, Internet of Things Journal
  • Reviewer / Program Committee for Conferences: ICLR 2025 , ICASSP 2025 , IJCNN 2025 , NAACL 2025 (Demo Track) , IJCAI 2025 (Demo Track), ICDCS 2024 (sub-reviewer) , EuroSys 2022 (Shadow PC) , IMC 2022 (Shadow PC)
  • Reviewer / Program Committee for Workshops: DL4C @ ICLR 2025, LLM Reason and Plan @ ICLR 2025, SCI-FM @ ICLR 2025, Multi-Agent AI in the Real World @ AAAI 2025 , ML and Compression @ NeurIPS 2024
  • Artifact Evaluation Committee: SIGCOMM 2021, SOSP 2021, MLSys 2023

  • Teaching
  • Teaching Fellow, CAS CS 210 Computer Systems, Boston University, Fall 2022
  • Teaching Assistant, CSCI 5105 Distributed Systems, University of Minnesota, Spring 2021
  • Teaching Assistant, CSCI 5103 Operating Systems, University of Minnesota, Fall 2020

  • Honors and Awards
  • OpenAI Researcher Access Program
  • Rank 16/183 in 2018 ACM-ICPC North Central North America Regional Contest
  • Bronze Medal, 2015 China Collegiate Programming Contest

  • Personal
    My reading notes of system papers.
    I love traveling and collecting old computer hardware. Here are the albums of my photography and collections.


    Using template from jonbarron.