Experiences
Education Experiences
-
Master of Science (Computer Science and Technology)
🏫 Fudan University, Sep. 2022 - Jun. 2025
-
Bachelor of Science (Elite Program in Computer Science and Technology)
🏫 Fudan University, Sep. 2018 - Jun. 2022
Industry Experiences
-
Platform Engineering Intern
☁️ Alibaba Cloud (China), Jun. 2021 - Sep. 2021
-
Explored the high-performance packet processing technique XDP for efficient network packet filtering.
-
Addressed real-world needs and further combined DPDK with XDP to handle network packets more intelligently.
-
Wrote and integrated technical documents to facilitate clear communications across teams
Research Experiences
-
Compress Lineage Artifacts in Lineage-Enabled DBMS [In progress]
🌐 Remote (Columbia University), Jun. 2024 - Present
Collaborated with Prof. Eugene Wu.
-
Explored the properties of various lineage artifacts across different input data, physical operators, and queries.
-
Designed a suitable compression strategy for each lineage artifact and integrated it into a lineage-enabled DBMS.
-
Evaluated on TPC-H benchmarks, achieving a ~6x space saving with an 18% additional time overhead.
-
Improve Tabular Data Quality for ML Models [2 submissions, 1 shepherd accepted by SIGMOD 2025]
🏢 Shenzhen Institute of Computing Sciences, Dec. 2023 - Oct. 2024
Collaborated with Prof. Wenfei Fan and Dr. Weilong Ren
-
Aimed to repair and enhance tabular data to improve ML classification performance across various metrics.
-
Identified key factors in tabular data impacting ML models and applied refined-scope data cleaning.
-
Designed iterative frameworks to address intractable cleaning problems and outperformed the SOTA by 30%.
-
Akane: Perplexity-Guided Time Series Data Cleaning [SIGMOD 2024]
🏫 Fudan University, Apr. 2023 - Feb. 2024
Collaborated with Prof. X. Sean Wang, Prof. Zhenying He and Prof. Peng Wang
-
Innovatively analogized recurrent patterns in the time series data to the word combinations in the textual data.
-
Re-formalized the time series data cleaning problem to minimize the perplexity of the time series.
-
Proposed a four-phase framework Akane and several advanced solutions to handle the NP-Hard cleaning problem.
-
Compared with 11 baselines across 12 real-world datasets with real errors and achieved SOTA performance.
-
RpDelta: Supporting UCR-Suite on Multi-Versioning Time Series Data
🏫 Fudan University, Mar. 2022 - Jan. 2023
Collaborated with Prof. X. Sean Wang and Prof. Zhenying He
-
Focused on the issues of multi-versioning time series data arising from various cleaning algorithms.
-
Designed an effective storage strategy RpDelta and saved over 5x space for multi-versioning time series data.
-
Transplanted the subsequence matching algorithm UCR-Suite onto RpDelta and doubled its speed cumulatively.
Contest Experiences
-
"Loongson Cup" National Student Computer System Capability Challenge [Second Prize]
🏫 Fudan University, Jun. 2020 - Aug. 2020
Collaborated with Prof. Liang Zhang and Dr. Chen Chen
-
Implemented an experimental dual-issued MIPS CPU & a five-stage pipelined MIPS CPU using System Verilog.
-
Incorporated memory hierarchy design for optimization (~30x faster) with great compatibility.
-
Burned the implemented CPU into an FPGA and initiated PMON on it.
-
Served as the leader of a four-person team, orchestrating project phases and refining task assignments.