Zhenyu Xu Email and Phone Number

Lab instructor for CS-2413: Data Structures at Texas Tech University (TTU), Lubbock, during Fall 2023, Spring 2024, Fall 2024 and Spring 2025. Facilitated hands-on learning experiences and guided students in implementing data structure concepts. Provided feedback on assignments and assisted students with debugging and code optimization.

-Optimized Small Language Models (SLMs) using GPTQ and SmoothQuant techniques via TensorRT and Neural Compressor. Reduced model size by approximately 60% with GPTQ-W8 quantization. Exported models using ONNX for edge deployment, achieving 1.38x inference speed improvement with SmoothQuant-W8A8 on specific tasks.-Developed Simple Eval framework for SLMs, incorporating metrics like PPL, token generation speed, and LWPA. Evaluated 6 open-source models (110M-1.7B parameters) on code generation tasks.-Analyzed code generation performance for 1M-350M parameter models. Tested various architectures. Created Gradio-based UI for model comparisons. Researched 10+ mainstream SLMs.