个人简介 PROFILE
计算机科学本科在读,已取得软件开发文凭,具备 C# / ASP.NET Core 与 Python 开发基础。做过 ASP.NET Core 业务系统,也做过用 LLM Agent 处理逾期发票风险判断与催收决策的流水线。曾在 EduGlobe 实习,搭建带人工审核的 AI 内容运营工作流。中英双语流利,正在寻找软件开发或 AI 应用方向的实习,可远程。
Computing Science undergraduate with a completed Diploma in Software Development and a working foundation in C# / ASP.NET Core and Python. My projects range from ASP.NET Core business applications to an LLM agent pipeline that assesses overdue-invoice risk and decides collection actions. Interned at EduGlobe, building an AI content operations workflow with human review. Fluent in English and Chinese. Seeking a software or AI application internship; open to remote.
实习经历 INTERNSHIP
EduGlobe
2026.05 - 2026.06AI 自动化与数字运营实习生 AI Automation & Digital Operations Intern
Tech: Python, PowerShell, SQLite, Obsidian, Buffer · GitHub
- 面向留学教育内容运营,将竞品研究、趋势采集、报告生成、文案与图片制作、人工审核和多平台发布准备,拆解为一条端到端的 AI 内容工作流。 Broke down study-abroad content operations into an end-to-end AI content workflow: competitor research, trend collection, reporting, copy and image production, human review and multi-platform publishing preparation.
- 使用 Python 与 PowerShell 构建趋势与竞品雷达,接入 RSS、Reddit、Apify 和手工 CSV 共 4 类输入,实现相关度评分、时效筛选、去重,以及受限数据源的降级处理。 Built a trend and competitor radar in Python and PowerShell over 4 input types (RSS, Reddit, Apify and manual CSV), with relevance scoring, recency filtering, deduplication and a fallback queue for restricted sources.
- 用 Obsidian 管理趋势卡片、竞品档案、研究报告和每日选题;通过 SQLite 内容状态和 Buffer 发布门禁,保证所有内容在发布前经过人工确认。 Organised trend cards, competitor profiles, reports and daily content choices in Obsidian; a SQLite content state and a Buffer publishing gate keep a human approval step before anything is published.
- 基于官方课程手册的事实生成招生海报与文案,审核图片统一为 1080 × 1350 px;用 JSONL 运行事件、健康检查和只读的 Supervisor 进程记录每次运行的完成与崩溃情况。 Generated enrolment posters and captions grounded in official course brochures, standardised to 1080 × 1350 px for review; recorded completed and crashed runs with JSONL events, health checks and an observation-only supervisor process.
项目经验 PROJECTS
智能账单催收引擎 Intelligent Billing Operations Engine
Tech: Python, Google Gemini API, Pydantic, pytest · GitHub
- 搭建三段式流水线:逾期发票 → 风险推理 → 催收策略 → 执行与审计,用 LLM 推理替代硬编码规则树,输出风险等级、置信度和判断理由。 Built a three-stage pipeline (risk reasoning, collection strategy, execution and audit) that replaces hardcoded rule trees with LLM reasoning and returns a risk level, confidence and reasons for each overdue invoice.
- 用 Pydantic 定义各阶段的输入输出契约,并在模型调用失败时回退到确定性规则,保证流程不中断。 Defined typed input/output contracts for every stage with Pydantic, with a deterministic rule fallback when a model call fails.
- 每一步决策写入结构化审计日志,催收动作通过模拟通道执行,并编写了单元测试与集成测试。 Logged every decision as a structured audit event; collection actions run through a simulated channel, covered by unit and integration tests.
智能证书管理系统 SmartCertificateSystem
Tech: ASP.NET Core MVC, C#, EF Core, SQLite, Razor, xUnit · GitHub
- 开发学校证书管理应用,覆盖管理员维护、学生成绩单与 GPA 查询、雇主验证,以及无需登录的公开证书验证 4 类业务路径。 Developed a school certificate management application covering 4 flows: administrator maintenance, student transcript and GPA lookup, employer verification, and public certificate verification without login.
- 基于 Session 与角色授权实现差异化访问和下载权限;系统没有存储文件时,根据数据库记录按需生成 PDF 成绩单。 Implemented session and role-based authorisation for access and downloads; generates PDF transcripts on demand from database records when no stored file exists.
- 编写 13 项 xUnit 测试覆盖主要业务行为,并整理演示账号、验证样例和运行说明,支持完整流程演示。 Wrote 13 xUnit tests covering the main business behaviour, and prepared demo accounts, verification samples and run instructions for a full walkthrough.
蒸好味收银台 Point-of-Sale Web App
Tech: JavaScript, Cloudflare Pages Functions, D1 (SQLite)
- 为餐饮店开发单页收银应用,支持下单、作废和出餐状态,订单通过 Pages Functions 接口写入 D1 数据库并按日查询。 Built a single-page POS app for a food stall with order entry, voiding and completion status; orders are stored in D1 through a Pages Functions API and queried by day.
志愿经历 VOLUNTEER EXPERIENCE
ActiveSG (新加坡体育理事会) ActiveSG
赛事志愿者 Event Volunteer
- 参与大型体育赛事的现场组织与协调,负责人群引导及突发情况处理。 Supported on-site organisation and coordination of major sporting events, including crowd guidance and incident handling.