Ziying Yang
Full-Stack Engineer | AI & Multi-Agent Systems | Python, TypeScript
Education
University of New South Wales (UNSW), Sydney
2024 — 2026Master of Information Technology, Information Technology
Shanghai Lixin University of Accounting & Finance
2018 — 2022Bachelor of Management, Information Management and Systems
Experience
QCon ConferencesTechnical Volunteer -- Lecturer Assistant
2019 — 2019- Coordinated logistics for conferences with speakers from Baidu, Alibaba, Ant Group, and 4Paradigm.
- Translated complex AI topics into accessible presentation materials for diverse audiences.
Projects
SkyTerrain -- AI-Enhanced 3D Geospatial PlatformFull-Stack Application
2026 — Present- Built a production 3D geography platform using Next.js, TypeScript, and CesiumJS.
- Architected a unified data pipeline for terrain metadata and educational content.
- Developed a terrain-aware camera system generating observation points automatically.
- Integrated AI narration with bilingual Chinese/English support and sentence highlighting.
ADHD-Timebox -- Multi-Agent AI Automation PlatformAI Application
2026 — Present- Engineered a cross-platform AI app (Web + Electron) with FastAPI backend.
- Architected a multi-agent system with Orchestrator, Planner, Focus, and Reward agents.
- Integrated Google Calendar API via OAuth 2.0 for real-time task synchronization.
- Implemented reactive idle-watching and context relevance checking for focus management.
Multi-Agent Code Translation SystemResearch Project
2025 — 2025- Developed a multi-agent system for automated code translation across Python and Java.
- Designed a modular orchestration engine with planner, translator, and reviewer agents.
- Leveraged AST analysis, sandboxed execution, and iterative validation for reliability.
- Co-authored technical documentation supporting reproducibility and team onboarding.
Involution Hell -- Open-Source Developer PlatformOpen Source Project
2025 — 2025- Engineered an open-source learning platform using Next.js with MDX content management.
- Integrated multi-provider AI models: OpenAI GPT, Google Gemini, and Intern-AI.
- Resolved critical OAuth authentication issues for stable global Vercel deployment.
Multimodal RAG Chatbot -- Technical PDF Question AnsweringResearch Project
2025 — 2025- Designed a multimodal RAG system for question answering over technical PDFs.
- Built an end-to-end pipeline using MinerU to parse PDFs into structured elements.
- Implemented embedding-based retrieval with ChromaDB and metadata-aware grounding.
- Generated answers using Vision-Language Models with multimodal context construction.
Skills
Programming Languages:
Python, TypeScript, JavaScript, Java, C++, Bash
AI & Machine Learning:
Multi-Agent Systems, RAG, Conversational AI, LLM Orchestration, Prompt Engineering, Vision-Language Models, ChromaDB, Vector Databases
Frontend & Backend:
Next.js, React, Node.js, FastAPI, Tailwind CSS, Electron, RESTful APIs, OAuth
Cloud, DevOps & Tools:
AWS, Vercel, Docker, CI/CD, Git, Linux, CesiumJS
Languages:
English (Fluent), Mandarin Chinese (Native)