Curriculum Vitae

Ph.D. Candidate in Computational Mathematics

School of Mathematical Sciences, Fudan University

Research focus: quantum computing and fast algorithms. Advised by Prof. Yingzhou Li.

B.Sc. in Mathematics and Applied Mathematics

School of Mathematical Sciences, Fudan University

Mathematics Talent Program. Awarded Outstanding Graduate of Fudan University.

Kimi K3 Training Systems & Activation Offloading

Huawei, ICT BG Data Storage Product Line

• Independently built the full training software stack for Ascend 910C and 950 systems, including the operating system, drivers, runtime, and distributed-training dependencies; brought up Kimi K3 pruned-model pretraining with FSDP2 on a single-node A3 system with 8 Ascend 910C cards (16 dies, 8 × 2 × 64 GB).
• Evaluated activation offloading on an A5/Ascend 950 setup with high-speed interconnects and an external memory box; read and traced recomputation and offloading mechanisms in MindSpeed MM and selected activations for offloading at operator granularity.
• Used MS Insight to analyze forward hotspots and arithmetic intensity, and studied the amount of data that can be offloaded while overlapping computation, I/O, and communication.
• Analyzed HCCS communication paths and DMA behavior in multi-card communication and activation transfers, achieving about 15% end-to-end speedup over the recomputation baseline.

High-Frequency Futures Quantitative Research

Pengfu Tech

• Used statistical and machine learning methods to mine alpha factors from tick-level limit-order-book snapshot data.
• Worked with data from the Chinese commodity futures market to study high-frequency signals and quantitative research workflows.

Parallel PDE Solver

Scientific Computing Project

• Implemented a parallel Poisson–Boltzmann solver in C++ and MPI using the HIF-SI algorithm library.
• Ran scaling studies on the Tianhe-3 supercomputer and commercial cloud platforms, focusing on communication, workload distribution, and parallel efficiency.

Macro Placement Algorithm Research

Huawei Collaboration Project

• Developed and optimized C++ algorithms for 2D macro placement; evaluated results on industrial EDA workloads and merged code into Huawei's internal repository.
• Improved Open3DBench's 3D macro placement algorithm for placement quality and runtime efficiency.

HPC Cluster Administrator

Fudan University

• Built and maintain a multi-node HPC cluster, including networking, storage, user environments, and scientific software.
• Manage compute resources and jobs with Slurm; use Grafana and agent-assisted workflows for metrics collection, anomaly identification, and job analysis.