Skip to content
View moyuan10086's full-sized avatar

Block or report moyuan10086

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
moyuan10086/README.md

👨‍💻 Hi there, I'm Moyuan 👋

Typing SVG

Blog Email GitHub


🌟 About Me

  • 🎓 Education & Status:
    • B.Eng. in Cyberspace Security, Guangdong Polytechnic Normal University (GPNU) (2022.09 - 2026.07, Top 20%)
    • 🎯 Aiming higher & preparing for top-tier CS graduate studies while exploring LLM Systems & AI Security.
  • 🔬 Research & Work Experiences:
    • Peking University (PKU) · Data & AI Security Lab — AI Security Intern (JoySafeter Agent Evaluation Platform, Semgrep Sandbox, MCP/Skills Tooling)
    • Shenzhen Research Institute of Big Data (CUHK-SZ) — LLM AI Infra Intern (2026.06 - 2026.09) (Ascend 910C NPU Inference, Qwen-235B W8A8 Long-Context HSP Compression, vLLM-Ascend TP8+EP8 Distributed Parallelism)
    • NSFOCUS (绿盟科技) — Security Intern (Enterprise WAF/IDS Deployment, Vulnerability Assessment)
  • 🏆 Honors: National 1st Prize in 19th "Challenge Cup" AI Track (第十九届“挑战杯”全国一等奖) | National 3rd Prize in 19th CISCN & 3rd "Great Wall Cup" (第十九届全国大学生信息安全竞赛暨“长城杯”三等奖) | 1st Prize (Silver) in Greater Bay Area "AI for Science" Contest | ISCC 2024 National 2nd Prize

🛠 Tech Stack

Python PyTorch Ascend vLLM MCP FastAPI Docker PostgreSQL Redis Linux Git

  • LLM Systems & AI Infra: PyTorch, torch_npu, vLLM-Ascend, CANN/ATB, W8A8 Quantization, TP/EP Distributed Decoding, Flash Decoding, HSP Long-Context Compression
  • AI Safety & Red Teaming: garak, Prompt Injection & Jailbreak Defense, Semgrep Source Code Sandbox, Multi-modal Content Auditing, Input/Output Content Guardrails
  • Agent & Backend Engineering: MCP (Model Context Protocol) Servers, Agent Skills, FastAPI, WebSocket, Celery, ChromaDB, Keycloak, MinIO

🚀 Featured Projects

Project Description Stack
PeerAssist Evidence-driven AI academic reviewing workbench with PDF range citation checks, review timeline state machines, and immutable audit reports. FastAPI React PostgreSQL Keycloak MinIO
Unlimited-OCR-NPU Huawei Ascend 910 NPU adaptation and benchmarking for modern OCR models; achieved 100% parity with GPU accuracy on OCRBench STR. torch_npu vLLM-Ascend FastAPI
aigc-safety-system National 3rd Prize in 19th CISCN & "Great Wall Cup". Interpretable AIGC multimodal forgery detection & content auditing platform with 70,000+ redline knowledge base. CLIP PEFT ChromaDB FastAPI Vue3
digital-human-assistant Real-time multimodal AI digital human with end-to-end low-latency ASR, LLM dialogue generation, TTS, and lip-sync driving. FastAPI WebSocket Redis Volcengine
moyuan10086.github.io Personal tech blog & publication column featuring 50+ deep-dive technical articles on AI, systems security, and cryptography. Hexo Butterfly GitHub Pages

📊 GitHub Statistics

Moyuan's GitHub Stats Top Langs

Built with care · Welcome to connect and collaborate!

Pinned Loading

  1. moyuan10086.github.io moyuan10086.github.io Public

    HTML 1

  2. digital-human-assistant digital-human-assistant Public

    AI数字人对话系统 - 基于Wav2Lip、火山引擎的实时语音对话助手,支持会话管理

    Python 3