MPHIL SCIENTIFIC COMPUTING · ROBOTICS · AI FOR ENGINEERING
Zongmian Wu
University of Cambridge · Cavendish Laboratory
I work on performance-aware scientific computing and learning-enabled robotics. I hold the first-class BEng Electronic and Electrical Engineering degree at University College London.
GPU Acceleration for Scientific Workloads
MPhil · Scientific Computing (HPC)
CUDAPerformanceNumerics
- Focus on performance profiling, bottleneck analysis, and kernel-level optimisation.
- Aim for reproducible benchmarking and clean engineering documentation.
AI-driven TCAD Agent for Semiconductor Device Simulation
Software & AI Engineer · Hong Kong Science and Technology Park
RAGSimulationAgent Systems
- Developed an agent workflow for device simulation automation with retrieval + self-correction loops.
- Benchmarked retrieval strategies to improve accuracy and efficiency in engineering workflows.
Academic interests
High-performance computing (MPI/OpenMP/CUDA), scientific machine learning, robust perception-and-control pipelines in robotics, and AI-assisted engineering workflows for simulation and design.