Scofield Zhengqing Liu

📣 09/26 - DeterV won the Best Paper Award 🏆 at PLOS'26!

📣 07/26 - Remora is accepted to VLDB'26!

📣 06/26 - Joined Copper Robotics on building deterministic robotics!

📣 12/25 - SyncWise is accepted to NSDI'26!

About

I am a final-year Ph.D. student at Large-Scale Data & Systems Group (LSDS) of Imperial College London, advised by amazing Marios Kogias. I am broadly interested in computer systems and programming languages. My research explores determinism as a foundational primitive for building reliable and robust systems that are easier to debug, reproduce, and reason about. I apply this perspective across multiple domains, including datacenter microservices, blockchain smart contracts, distributed databases, systems verification, and robotics.

Prior to joining Imperial, I was fortunate to work with Yiting Xia (MPI-INF) on hardware-accelerated datacenter networks, Marc Shapiro (INRIA) on actor-based distributed query processing, and Romain Jacob (ETHz) on P4-based failure-rerouting. I obtained M.S. from École Polytechnique and B.S. from South China University of Technology.

Determinism stack: Runtime (DORADD, Remora, copper-rs); Verification (DeterV); Compiler (DeterCL) — Projects highlight: my determinism stack.

Selected Publications

DeterV: Verifying Deterministic Parallel Execution
Zhengqing Liu, Marios Kogias
14th Workshop on Programming Languages and Operating Systems (PLOS'26, colocated at SOSP'26)
🏆 Best Paper Award
Paper | Slides | DOI

Remora: Scale-out Deterministic Execution for Smart Contracts
Zhengqing Liu, Alberto Sonnino, Igor Zablotchi, Lefteris Kogias, Marios Kogias
International Conference on Very Large Data Bases (VLDB'26)
Supported by a Sui Academic Research Award and Mysten Labs
The 1st blockchain execution layer matching SOTA consensus performance
Paper | VLDB | Slides | DOI | Code

DORADD: Deterministic Parallel Execution in the Era of Microsecond-Scale Computing
Zhengqing Liu, Musa Unal, Matthew J. Parkinson, Marios Kogias
ACM Symposium on Principles & Practice of Parallel Programming (PPoPP'25)
In collaboration with Microsoft Research (Verona team)
300x latency and 2.5x throughput improvement than SOTA
Invited talks at Microsoft Research Cambridge and Subzero Labs
Paper | Slides | DOI | Code