Scholar
Chunwei Liu
Google Scholar ID: Q0LOhAgAAAAJ
Massachusetts Institute of Technology
Citations & Impact
Citations
642
H-index
13
i10-index
15
Publications
20
Co-authors
14
list available
Publications
Resume (English only)
Academic Achievements
- June 2025: Paper on cloud analytics workload synthesis tool, “PBench: Workload Synthesizer with Real Statistics for Cloud Analytics Benchmarking,” was accepted by VLDB 2025. Same month, another paper on non-intrusive DBMS scheduling, “Improving DBMS Scheduling Decisions with Accurate Performance Prediction on Concurrent Queries,” was also accepted by VLDB 2025. Additionally, a comprehensive study of lossless floating-point compression, “Beyond Compression: A Comprehensive Evaluation of Lossless Floating-Point Compression,” was accepted by VLDB 2025. March 2025: Demo paper on a chat interface for declarative AI frameworks, “PalimpChat: Declarative and Interactive AI Analytics,” was accepted by SIGMOD 2025. February 2025: Papers on scientific discovery and open columnar formats evaluation were accepted by AISD @ NAACL 2025 and the VLDB Journal special issue “Best of VLDB,” respectively. January 2025: Paper on LLM scheduling, “Don't Stop Me Now: Embedding Based Scheduling for LLMs,” was accepted to ICLR 2025.
Research Experience
- Currently a Postdoctoral Associate at MIT CSAIL, working with Michael Cafarella.
Education
- Received his Ph.D. from the Department of Computer Science at the University of Chicago, where he worked in the ChiData group advised by Aaron Elmore.
Background
- His research interests span compound AI systems, database systems, cloud/edge computing, and database benchmarking. Focuses on optimizing data systems for both conventional data analytics and emerging AI-powered pipelines. Develops privacy-preserving workload generation techniques for evaluating cloud database systems, collaborating with major cloud vendors such as Microsoft, Amazon, Intel, Meta, and Google. Also explores novel data compression methods, implementing adaptive compression selection in both conventional and resource-constrained databases and machine learning systems. Engaged in high-dimensional data analysis, with a particular emphasis on time series applications.
