👋 About Me

Pengcheng Zhang is a senior engineer and researcher at Tencent Big Data. His work focuses on efficient, scalable data infrastructure for multimodal AI workloads on heterogeneous hardware. Currently, he works on TEngine, a general-purpose, high-performance execution engine that powers TEngineDB (MPP) and TEngine Spark (BSP) across CPUs, GPUs, and NPUs.

He received his Ph.D. from Shanghai Jiao Tong University, under the supervision of Prof. Bin Yao.

His research spans query execution, distributed data processing, and vector search, with the resulting systems deployed at scale within Tencent, serving ad-hoc analytics, batch-processing, and vector-search workloads across Hy LLMs, social platforms, gaming, advertising, financial services, and e-commerce. His work has appeared in VLDB, SIGMOD, DSN, The VLDB Journal, and IEEE Transactions on Computers. He was selected for the Tencent Qingyun Plan and received the Alibaba Outstanding Academic Collaboration Project Award.

张鹏程聚焦新一代多模态 AI 负载的异构计算需求，致力于构建统一、高效、可扩展的数据计算基础设施。现任腾讯大数据高级工程师，从事通用异构计算引擎 TEngine 的研发。作为通用计算底座，TEngine 通过软硬件协同优化，为 TEngineDB（MPP）和 TEngine Spark（BSP）提供跨 CPU、GPU 和 NPU 的统一高性能执行能力。

博士毕业于上海交通大学，师从姚斌教授。

他的研究涵盖查询执行、分布式数据处理和向量检索，相关创新成果已应用于腾讯大数据基础设施并实现大规模部署，承载混元大模型、社交、游戏、广告、金融和电商等场景下的交互式分析、批处理与向量检索负载。相关工作发表于 VLDB、SIGMOD、DSN、The VLDB Journal 和 IEEE Transactions on Computers。入选腾讯青云计划，曾获阿里巴巴优秀学术合作项目奖。

For internship opportunities at Tencent or research collaboration, feel free to get in touch.

🔥 News

2026.06: TEngineDB-V, an OLAP-native vector search system for large-k workloads at Tencent, was accepted by VLDB 2026.
2026.06: TRIM was presented at ACM SIGMOD 2026.
2024.12: UNIFY was accepted by VLDB 2025.
2024.10: Gave a StarRocks × Tencent webinar on vector similarity search.

📝 Selected Publications

* Denotes the corresponding author.
^ Denotes equal contribution.

VLDB 2026 Xufei Wu^{^}, Pengcheng Zhang^{^}, Yitong Song, Xiaobo Zhang, Anqi Liang, Kai Wang, Jijun Du, Yidi Xiong, Guangxu Cheng, Zhe Chen, Peng Chen, Guoliang Li, Xuanhe Zhou, Fan Wu.
TEngineDB-V: An OLAP-Native Vector Search System for Large-k Workloads at Tencent. Proceedings of the VLDB Endowment, to appear, 2026. arXiv Code
IEEE TC 2026 Yaoguang Yong, Xiaoming Du, Yunong Wang, Pengcheng Zhang, Xuhua Ma, Bin Yao, Xudong Zheng, Linquan Jiang.
Collaborative Prediction of Cloud DRAM Failures with Rules and Machine Learning. IEEE Transactions on Computers, 75(4): 1622–1635, 2026. Link
SIGMOD 2026 Yitong Song, Pengcheng Zhang, Chao Gao, Bin Yao, Kai Wang, Zongyuan Wu, Lin Qu.
TRIM: Accelerating High-Dimensional Vector Similarity Search with Enhanced Triangle-Inequality-Based Pruning. Proceedings of the ACM on Management of Data, 3(6): 1–26, 2025. Pdf Code
DSE 2025 Jinguo You, Yuxuan Wang, Xingrui Huang, Zhenrui Yi, Wanting Fu, Kaiqi Liu, Pengcheng Zhang, Bin Yao.
SOC: A Succinct Adaptive Semantic OLAP Caching. Data Science and Engineering, 10(4): 621–638, 2025. Pdf Code
VLDB 2025 Anqi Liang, Pengcheng Zhang, Bin Yao, Zhongpu Chen, Yitong Song, Guangxu Cheng.
UNIFY: Unified Index for Range Filtered Approximate Nearest Neighbors Search. Proceedings of the VLDB Endowment, 18(4): 1118–1130, 2024. Pdf Code
VLDBJ 2023 Pengcheng Zhang, Bin Yao, Chao Gao, Bin Wu, Xiao He, Feifei Li, Yuanfei Lu, Chaoqun Zhan, Feilong Tang.
Learning-Based Query Optimization for Multi-Probe Approximate Nearest Neighbor Search. The VLDB Journal, 32(3): 623–645, 2023. Link
DSN 2022 Pengcheng Zhang, Yunong Wang, Xuhua Ma, Yaoheng Xu, Bin Yao, Xudong Zheng, Linquan Jiang.
Predicting DRAM-Caused Node Unavailability in Hyper-Scale Clouds. IEEE/IFIP International Conference on Dependable Systems and Networks, 275–286, 2022. Pdf

🏆 Honors & Awards

2024.07, Selected for the Tencent Qingyun Plan.
2023, Alibaba Outstanding Academic Collaboration Project Award, for the Alibaba Cloud ECS AIOps system.
2015, MATLAB Innovation Award, China Undergraduate Mathematical Contest in Modeling (awarded to 1 of 25,558 teams).
2015, National First Prize, China Undergraduate Mathematical Contest in Modeling.

💼 Experience

2024.07 – Present, Senior Engineer, Tencent Big Data. Working on TEngine, TEngineDB, and TEngine Spark across CPUs, GPUs, and NPUs.
2023.03 – 2024.07, Intern, Tencent Big Data. Led the development of TenANN and contributed vector search capabilities to StarRocks, a leading open-source OLAP database widely used by Airbnb, Alibaba, Tencent, and Xiaohongshu.
2021.06 – 2023.03, Intern, Alibaba Cloud. Worked on XBrainM, an AIOps system that has run in the Alibaba Cloud ECS production environment for years, supporting intelligent operations across more than one million nodes and reducing DRAM-caused node unavailability by 60%. This work resulted in a first-author paper at DSN 2022.

🎓 Education

2017.09 – 2024.06, Shanghai Jiao Tong University, Ph.D. in Computer Science and Engineering.
2013.09 – 2017.06, Xidian University, Bachelor’s degree in Information Security (Experimental Program).

📜 Selected Patents

授权王雨农、张鹏程、杜小明、马旭华。
特征提取方法、装置、电子设备及存储介质。中国发明专利 CN116127292B，2026（申请号 CN202211543083.X）。
申请 张鹏程、程广旭、蒋杰、刘煜宏、陈鹏、罗韩梅、杜佶峻、赵裕隆。
数据查询方法及其相关产品。中国发明专利申请 CN202411255378.6（公开号 CN121636545A），2026。
申请 张鹏程、程广旭、陈奕安。
数据处理方法、装置、设备、存储介质和产品。中国发明专利申请 CN202410039737.8（公开号 CN120296039A），2025。
授权徐姚亨、姚斌、张鹏程、唐飞龙、沈耀、郑文立。
一种基于Spark的大规模高维数据近似近邻查询系统和方法。中国发明专利 CN114329094B，2024（申请号 CN202111672312.3）。
授权陆元飞、魏闯先、楼仁杰、吴斌、姚斌、张鹏程、高超。
一种向量检索方法及装置。中国发明专利 CN113449132B，2022（申请号 CN202110991057.2）。

Pengcheng Zhang 张鹏程