University of California, Riverside
Ph.D student, Department of Computer Science
Co-advised by Prof. Yan Gu and Prof. Yihan Sun
Last updated in 2026/07
Xiangyun Ding
Education
Tsinghua University
Bachelor of Engineering, Department of Computer Science and Technology
Adviser: Prof. Wenjian Yu
Skills
Languages: C++, Python, Go, Java
AI/LLM: PyTorch, CUDA/ROCm programming, Triton/Tilelang
Awards and Honors
- 2025 Laxmi Bhuyan Fellowship, UC Riverside (3 recipients in the CS department each year)
- 2025 5th Place in SIGMOD 2025 Programming Contest 🔗
- 2024 Outstanding Paper Award, ACM Symposium on Parallelism in Algorithms and Architectures (SPAA)
- 2023 Best Paper Award, European Symposium on Algorithms (ESA)
- 2023 Dean’s Distinguished Fellowship, UC Riverside
- 2015 Silver Medal, Chinese 32nd National Olympiad in Informatics (CCF NOI 2015)
Work Experience
Software Engineer Intern
CasualFlow Inc.
- Optimize the LLVM AMDGPU backend. The new instruction scheduling algorithm gives up to 1.2x performance gain.
- Develop high-performance GPU kernels for LLM inference using CUDA/Triton/Tilelang, including GEMM, FlashAttention/KDA, Fused MoE/MegaMoE, in bf16/fp8/mxfp4 weights.
- Develop In-Context Reinforcement Learning (ICRL) for LLM-driven program optimization.
Software Engineer
Airbnb, Inc.
- Host tiering project backend development. Process data with airflow. Develop Java backend services.
- Airbnb chat-bot development. Build new bot rules for user-bot interaction. Use ML models to direct bot behaviors.
- Coding Interviewer.
Software Engineer Intern
- Processed and analyzed large-scale Wikipedia revision histories for the WikiTrust vandalism detection project.
- Designed and implemented rule-based feature extraction and heuristic classifiers.
Publications
Preprints
-
2026
ARGUS: Agentic GPU Optimization Guided by Data-Flow Invariants 🔗
Haohui Mai, Xiaoyan Guo, Xiangyun Ding, Daifeng Li, Qiuchu Yu, Chenzhun Guo, Cong Wang, Jiacheng Zhao, Christos Kozyrakis, Binhang Yuan.
Accepted for ACM SIGOPS Symposium on Operating Systems Principles (SOSP), 2026.
Publications
-
2026
Parallel Metric Skip Lists and Nearest Neighbor Search 🔗
Xiangyun Ding, Rohin Garg, Yan Gu, Yihan Sun. (Authors listed alphabetically)
In 38th ACM Symposium on Parallelism in Algorithms and Architectures (SPAA), 2026.
-
2025
New Algorithms for Incremental Minimum Spanning Trees and Temporal Graph Applications 🔗
Xiangyun Ding, Yan Gu, Yihan Sun.
In SIAM Conference on Applied and Computational Discrete Algorithms (ACDA), 2025.
-
2024
Parallel and (Nearly) Work-Efficient Dynamic Programming 🔗
Xiangyun Ding, Yan Gu, Yihan Sun.
In 36th ACM Symposium on Parallelism in Algorithms and Architectures (SPAA), 2024. Outstanding paper award.
-
2024
Fast and Space-Efficient Parallel Algorithms for Influence Maximization 🔗
Letong Wang, Xiangyun Ding, Yan Gu, Yihan Sun.
In Proceedings of the VLDB Endowment (VLDB), 2024.
-
2023
Efficient Parallel Output-Sensitive Edit Distance 🔗
Xiangyun Ding, Xiaojun Dong, Yan Gu, Youzhe Liu, Yihan Sun. (Authors listed alphabetically)
In 31st Annual European Symposium on Algorithms (ESA), 2023. Best paper award.
-
2022
DP-Nets: Dynamic programming assisted quantization schemes for DNN compression and acceleration 🔗
Dingcheng Yang, Wenjian Yu, Xiangyun Ding, Ao Zhou, Xiaoyi Wang.
In Integration, the VLSI Journal, 2022.
-
2020
Efficient model-based collaborative filtering with fast adaptive PCA 🔗
Xiangyun Ding, Wenjian Yu, Yuyang Xie, Shenghua Liu
In IEEE 32nd International Conference on Tools with Artificial Intelligence (ICTAI), 2020.
Other Projects
LQ-Nets-PyTorch 🔗
- A PyTorch implementation of LQ-Nets (neural network quantization).
Parallel-GPT-2 🔗
- A C++ implemented GPT-2, running on multi-CPUs in parallel.
- Up to 3x faster than llm.c on 64 cores.
Vahadane: Histological Image Normalization 🔗 (80+ stars)
Single-user Relational Database 🔗
Five-level Pipelined CPU 🔗
- A five-level pipelined CPU implementation on FPGA using Verilog and the MIPS instruction set.
Distributed Sharded Key/Value Storage 🔗
- A distributed KV database written in Go. Implemented the Raft consensus algorithm.