Monai dehudewf

About Me

🧠 Passionate about LLM Inference Optimization / Computer Vision / AI Systems

🎓 CS Undergrad → HK MSc in AI (2026)
My tech blog https://grape.mobi/
Working on LLM inference optimization — vLLM / KsanaLLM / CUDA
Research in Multimodal medical image fusion & object detection (YOLOv10, GCN, Mamba)
🥉 Kaggle Medalist — BYU Bacterial Flagellar Motors (top 9%, Bronze)
Ask me about LLM Serving, CUDA Optimization, vLLM, Computer Vision, or Kaggle
Fun fact: I read vLLM source code for fun and reduced KV Cache memory by 98.3% ⚡

Find me around the web:

Tech Stack & Tools

class Monai:
    username = "dehudewf"
    education = {"undergrad": "CS (top 5%)",
                 "master": "HK MSc — Artificial Intelligence (2026)"}
    languages = ["Python", "C++", "C", "CUDA", "Java", "TypeScript", "SQL", "LaTeX"]
    ai_stack = {"llm_inference": ["vLLM", "KsanaLLM", "FlashAttention", "PagedAttention"],
                "computer_vision": ["PyTorch", "OpenCV", "YOLOv10", "GCN", "Mamba"],
                "data_science": ["NumPy", "Pandas", "scikit-learn", "Jupyter"]}
    tools = {"editors": ["Cursor", "VS Code", "Neovim"],
             "devops": ["Git", "Docker", "Linux"],
             "databases": ["PostgreSQL", "MySQL"]}
    current_focus = "LLM Inference Optimization (vLLM/CUDA) & Multimodal CV Research"
    fun_fact = "Reduced KV Cache cross-machine communication latency by 98.3%"

GitHub Stats

Featured Projects

🚀 Project	📝 Description	🛠️ Tech
🧬 Kaggle BYU - Bacterial Flagellar Motors	Kaggle 竞赛：定位细菌鞭毛马达
🤖 LLM Prompt Recovery	LLM 提示词恢复项目
🏛️ Suzhou Pavilion - OpenGL	基于 OpenGL 的苏州白鹭园亭 3D 建模
🎬 Movie Data Analysis	电影数据分析与可视化