- M.Eng. Software Engineering student at USTC (University of Science and Technology of China), Hefei/Suzhou, China
- Researching LLM behavior in offensive security at USTC's Information Security Lab - lead developer on Pentra-AI, an MCP-based pentest framework built on top of an open-source base called Hexstrike-AI
- Background in systems programming, network security, and full-stack dev
- French national based in China, working across the CN/EU tech stack
- Coursework: Information Security, Compiler Engineering, Blockchain, Information Theory & Coding, HarmonyOS Development
Pentra-AI — LLM Security Research Framework
Python/Flask · MCP · Kali Linux · Anthropic & Deepseek APIs
Research project at USTC's Information Security Lab. Extended a pentest framework exposing 150+ real security tools via Model Context Protocol, investigating LLM tool-selection behavior across CTF challenges. Compared Claude vs DeepSeek across 3 AI clients using a 774-prompt benchmark corpus.
Stardew Dashboard — Full-Stack AI Dashboard
PHP
Intelligent dashboard analyzing Stardew Valley save files.
Unity · C# · AI State Machines
Team-built game with enemy AI using state machines and pathfinding. Awarded 3rd place nationally at Festival MMI France.
