Feiyang Zheng

zhengfeiyang

📍 Raleigh, NC, United States
📧 fzheng4@ncsu.edu
🔗 LinkedIn
🌐 GitHub Profile
📄 Resume (PDF)

💡 About Me

I am a Computer Science PhD student at North Carolina State University, co-advised by Dr. Frank Mueller and Dr. Jiajia Li. My research sits at the intersection of high-performance computing, scientific data compression, and GPU systems.

Previously, I completed my B.S. in Mathematics and Applied Mathematics at Southern University of Science and Technology (SUSTech), where I gained hands-on experience in supercomputing competitions, parallel programming, and large-scale AI model optimization.

I am interested in bridging applied mathematics and system-level design — building faster, smarter computing systems through innovations in tensor computation, floating-point representation, and parallel frameworks.

🔬 Current Research

Scientific Data Compression Without Decompression

C++, CUDA, Python (Aug 2025 – Present, ongoing)

Developing lossy compression for large-scale scientific simulation data using low-rank and quantization-based methods, designed so that sub-regions can be accessed and processed directly on the compressed representation without full decompression.
Implemented the compression pipeline in C++/CUDA, with Python tooling for parameter sweeps and evaluation on public scientific datasets.
Part of NSF EPOD: Data Reduction for Science through Efficient Processing Without Decompression (award #2504512).

GPU Kernel Data-Race Detector

C++, Python, CUDA (2026 – Present, ongoing)

Building a data-race detector for CUDA kernels (unpublished; details available on request).

🎓 Education

North Carolina State University, Raleigh, NC

Ph.D. in Computer Science (Aug 2025 – May 2030, expected) · GPA 3.75/4.3 Co-advised by Dr. Frank Mueller and Dr. Jiajia Li.

Graduate coursework: Architecture of Parallel Computers (CSC/ECE 506) · Advanced Computer Architecture (ECE 786) · Compiler Construction (CSC 512) · Code Optimization (CSC 766) · Numerical Analysis I (CSC 580)* · Software Engineering (CSC 510)* _{* in progress}

Southern University of Science and Technology (SUSTech), Shenzhen, China

Bachelor of Science in Mathematics and Applied Mathematics (Sep 2021 – Jun 2025)

Selected CS coursework: Operating Systems · Computer Organization · Digital Logic · Computer Networks · Data Structures and Algorithm Analysis (Honors) · Principles of Database Systems · Machine Learning · Scientific Computing

📄 Publications

Dogan Corus, Pietro S. Oliveto, Feiyang Zheng. “Hybrid Selection Allows Steady-State Evolutionary Algorithms to Control the Selective Pressure in Multimodal Optimisation.” Genetic and Evolutionary Computation Conference (GECCO), 2025. DOI: 10.1145/3712256.3726411
Developed hybrid selection mechanisms to improve escape from local optima and better control selective pressure in multimodal fitness landscapes. Experiments in C++, analysis and visualization in MATLAB.

🧑‍🏫 Teaching

Teaching Assistant — CSC/ECE 506: Architecture of Parallel Computers

North Carolina State University, Summer 2026

Authored the programming assignments (parallel architecture, shared-memory programming) and a portion of the exam problems.
Graded assignments and supported students.

Teaching Assistant — CSC 514: Foundations of Cryptography

North Carolina State University, Fall 2026

🧾 Service

Sub-reviewer: ICS 2026 · SSDBM 2026 · SC 2026

💼 Prior Experience

Research Assistant — Southern University of Science and Technology

(Sep 2024 – Jun 2025)

Built a LibTorch-based deep learning package for modeling unknown equations.
Structured the project using CMake, integrating CUDA and yaml-cpp.
Designed C++ templates for flexible model construction and secondary development.

🧠 Earlier Projects (Undergraduate)

🏆 ASC22-23 Student Supercomputer Challenge — First Prize & Group Competition Award

Python, TensorFlow, DeepMD, Horovod, Spack (2022 – 2023) · announcement

Trained DeepMD, a deep-learning interatomic potential model, and ran molecular dynamics with it.
Configured a 4-node, 8×A100 cluster from scratch using Spack and Conda.
Applied the TensorFlow Profiler to identify optimization opportunities.

🏅 6th Asia-Pacific HPC-AI Student Competition — Second Place

Python, PyTorch, Nsight (2023) · results

Optimized BLOOM large-language-model inference for multi-node parallelism.
Profiled and improved efficiency using Nsight and the PyTorch Profiler.

⚙️ GEMM Implementations

C++, CUDA, OpenMP, SIMD, cuBLAS, VTune (Feb 2024 – Jul 2024)

Implemented GEMM from scratch in C++.
Optimized the CPU path with OpenMP and SIMD; located bottlenecks with VTune.
Wrote custom CUDA kernels and benchmarked them against cuBLAS.

💻 Technical Skills

Languages: C/C++, Python, Java, MATLAB, SQL (Postgres)
Parallel & GPU: CUDA, OpenMP, SIMD/AVX, PTX/SASS
Libraries: cuBLAS, LibTorch, PyTorch, TensorFlow, Horovod
Profiling & Analysis: Nsight Systems/Compute, VTune, TensorFlow Profiler, nvdisasm, cuobjdump
Developer Tools: Git, CMake, Docker, Spack, Conda, LaTeX

🤝 Let’s Connect

I am looking for a Summer 2027 internship in high-performance computing, GPU systems, compilers, or ML infrastructure. Feel free to reach out:

📧 fzheng4@ncsu.edu

🔗 LinkedIn

🌐 GitHub