Zhuoyang WuSYSU · AI
中山大学网络空间安全学院大三本科生。专注Ai Agent，并探索具身智能（VLA 模型与世界模型）， 探索机器人如何感知物理世界并自主决策，目前还是玩机械臂阶段。
I'm a third-year undergraduate at Sun Yat-sen University's School of Cyber Science and Technology. My work centers on AI agents, with a growing interest inembodied intelligence, especially VLA models and world models. Right now, that mostly means learning by building and experimenting with robot arms.
中山大学网络空间安全学院大三本科生（2027 届）， 就读于深圳校区，GPA 3.66/5.0。拥有约一年半的科研经验，主要方向为大模型训练数据去重和 AI 基础设施。就在7月1号，我考完了大学期间最后一门考试，结束快乐！我的论文投到ICDE还在审稿，希望一切顺利，审稿人求放过。
此前在深信服科技实习，负责自进化项目的数据挖掘算法开发。这是一个艰巨的项目！ 目前在探索具身智能——VLA 模型与世界模型。 最新在做我的论文阅读项目，我认为源源不断输入论文是很重要的，就算ai再强，我们也要丰富自己，做最后一环的决策，做一个适合大众并拥抱现在ai的阅读项目！
"做一点有意义的工作"
I'm a third-year student at Sun Yat-sen University's School of Cyber Science and Technology, based on the Shenzhen campus and graduating in 2027. My GPA is 3.66/5.0, and I've spent the past year and a half working on training-data deduplication and AI infrastructure. I wrapped up my final exam of the academic year on July 1, and my ICDE submission is still under review. Fingers crossed.
Before that, I interned at Sangfor Technologies, where I developed data-mining algorithms for a self-evolving system. It was a demanding project and a valuable lesson in turning research ideas into systems that have to work. I'm now exploring embodied intelligence through VLA models and world models. I'm also building a paper-reading project around a simple belief: even as AI becomes more capable, people still need to keep reading, form their own judgment, and remain the final link in the decision-making chain.
"Do work that matters."
一个面向文本、图像、音频及图文对的数据清洗研究项目，重点区分候选召回、重复关系判断与安全删除，并为实验结果保留可追溯证据。A research system for cleaning text, image, audio, and image-caption data, separating candidate retrieval, duplicate relation typing, and safe deletion with traceable experiment evidence.
为终端安全数据构建语义与行为图双维度的增量聚类管道，处理生长告警、簇心更新和运营 review 包交付。An incremental clustering pipeline for endpoint-security data, combining semantic and behavior-graph signals with growth-event deduplication and review-package delivery.
一个学习阶段的 Python Agent 框架：实现统一 LLM 接口、四类 Agent、工具与 MCP、分层记忆和 RAG，也诚实记录尚未解决的路由与自修正问题。A learning-stage Python agent framework implementing a unified LLM interface, four agent types, tools and MCP, layered memory, and RAG—along with its unresolved routing and self-correction limits.
一个由 Go、Python、Redis Stream 和 PostgreSQL 组成的个人信息过滤系统：抓取 RSS，去重并评估内容，再通过网页和 Telegram 交互。A personal information-filtering system built with Go, Python, Redis Streams, and PostgreSQL for RSS ingestion, deduplication, LLM evaluation, and web or Telegram interaction.
从真机尺寸、腕部视角和版本隔离开始，记录 SO-101 蓝色胶带任务的数字孪生搭建，以及一次没有被包装成成功的抓取诊断。Building a digital twin of the SO-101 tape task from real measurements and wrist views, then diagnosing a grasp that still failed the physics gate.
记录从 50 条真机示教、RTX 3090 远程推理，到三种动作交接路线对比的 π0.5 真机闭环。A practical record of 50 real-robot demonstrations, remote π0.5 inference on an RTX 3090, and three action-handoff strategies.
记录我从采集第一批示教数据，到训练 ACT、排查动作块问题，最终让 SO-101 在固定场景完成抓放任务的过程。A practical record of collecting demonstrations, training ACT, debugging action chunks, and closing the loop on a real SO-101 robot.
记录 SO-101 主从机械臂从舵机准备、结构装配到上电检查的过程，以及几个最容易返工的地方。A hands-on record of preparing the servos, assembling the leader and follower arms, and checking the hardware before power-up.
7月21日对前半个月的阶段性总结A mid-July check-in on IELTS, research, projects, and what comes next.
秋招准备计划：时间线、项目筹备、基础知识复习、算法训练、面试准备。My plan for projects, core CS review, algorithms, and interview preparation.
第一篇文章 — 关于我为什么开始写作。My first essay on why I want to start writing again.