I work at the AI lab of a private quantitative hedge fund, focusing on AI infrastructure and ML systems. My research interests center on AI agents for long-horizon tasks such as agentic coding and autonomous research workflows.

Before that, I was a member of Team Viking at ByteDance (2024–2026), working on the Viking engine — an approximate nearest neighbor search engine that drives recommendation systems for TikTok and Douyin, and also powers VikingDB, the vector database of Volcano Engine.

I joined Moqi as its first engineer (2017–2024) and led the infrastructure team. There, I directed the development of MyScale, a high-performance vector and full-text search engine based on ClickHouse. I also built two nationwide billion-scale Automated Fingerprint Identification Systems (AFIS) for the Chinese public security, and created the world’s first contactless fingerprint scanner, the Moqi A1, which achieves image quality comparable to traditional contact-based scanners.

Earlier, I worked as a researcher at Huawei Noah’s Ark Lab (2016–2017).

I hold a PhD in Computer Science from The Chinese University of Hong Kong (CUHK, 2016), supervised by Prof. John C.S. Lui at the Advanced Networking and System Research Laboratory (ANSRLab), where I focused on graph analytics and stream processing systems. I earned my Bachelor’s degree from Shanghai Jiao Tong University (ACM Honors Class).

Publications

Conference:

• Graph Edge Partitioning via Neighborhood Heuristic
  Chenzi Zhang¹, Fan Wei¹, Qin Liu, Zhihao Gavin Tang, Zhenguo Li.
  The 23rd SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2017 (oral presentation)
  [paper] [code]
  (AR of papers with oral presentation: 64/748 = 8.6%)
• PowerWalk: Scalable Personalized PageRank via Random Walks with Vertex-Centric Decomposition
  Qin Liu, Zhenguo Li, John C.S. Lui, Jiefeng Cheng.
  The 25th ACM International Conference on Information and Knowledge Management (CIKM), 2016 (long paper)
  [paper] [slides] [techreport] [code]
  (AR of long papers: 165/935 = 17.6%)
• Walking in the Cloud: Parallel SimRank at Scale
  Zhenguo Li, Yixiang Fang, Qin Liu, Jiefeng Cheng, Reynold Cheng, John C.S. Lui.
  Proceedings of the VLDB Endowment (PVLDB), 9(1), 2015 (also in VLDB 2016)
  [paper] [poster on SoCC'15]
• VENUS: Vertex-Centric Streamlined Graph Computation on a Single PC
  Jiefeng Cheng, Qin Liu, Zhenguo Li, Wei Fan, John C.S. Lui, Cheng He.
  The 31st IEEE International Conference on Data Engineering (ICDE), 2015
  [paper] [slides] [poster]
• SAND: A Fault-Tolerant Streaming Architecture for Network Traffic Analytics
  Qin Liu, John C.S. Lui, Cheng He, Lujia Pan, Wei Fan, Yunlong Shi.
  The 44th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN), 2014 (practical experience report)
  [paper] [slides]
  (AR: 56/185 = 30.3%)
• Combining Factorization Model and Additive Forest for Collaborative Followee Recommendation
  Tianqi Chen, Linpeng Tang, Qin Liu, Diyi Yang, Saining Xie, Xuezhi Cao, Chunyang Wu, Enpeng Yao, Zhengyang Liu, Zhansheng Jiang, Cheng Chen, Weihao Kong, Yong Yu.
  KDD-Cup Workshop, 2012 (First place in Track 1)
  [paper] [slides]

Journal:

• VENUS: A System for Streamlined Graph Computation on a Single PC
  Qin Liu, Jiefeng Cheng, Zhenguo Li, John C.S. Lui. IEEE Transactions on Knowledge and Data Engineering (TKDE), 28(8): 2230-2245, 2016
  (An earlier version appeared in ICDE 2015)
  [paper] [doi]
• SAND: A fault-tolerant streaming architecture for network traffic analytics
  Qin Liu, John C.S. Lui, Cheng He, Lujia Pan, Wei Fan, Yunlong Shi.
  Journal of Systems and Software (JSS), 122: 553-563, 2016
  (An earlier version appeared in DSN 2014)
  [paper] [doi]