Full-Stack Developer & DevOps Specialist: 5+ years of experience delivering 20+ high-impact projects for industry leaders including Tesla, Nissan, YanFeng, and CASIC. Expert in building scalable cross-platform applications with end-to-end DevOps support and proven excellence in technical leadership.
AI Engineer & Researcher: Experience spans reinforcement learning, fine-tuning, synthetic data construction, and AI agent building. Passionate about trying the latest SOTA AI techs and building useful AI solutions. An active practitioner with continual $200/month GPT/Claude subscriptions and over $1,500 monthly token consumption.
AI: Context Engineering, MCP, Skills, RAG, ACP, PyTorch, LangGraph, Vertex AI, Agent Memory, Groq, Ollama, RL, Knowledge Distillation Frontend: TypeScript, React, Next.js, Vue, Tailwind, PWA, Micro-frontend, Capacitor.js Backend: FastAPI, NestJS, SpringCloud, PostgreSQL, MongoDB, Redis, Prisma, RabbitMQ 3D: Three.js, Spline, WebGL/webGPU, Unity, Godot, Tripo AI DevOps: AWS, GCP, Cloudflare, Supabase, Firebase, Docker, Kubernetes, Github Actions
I designed and developed an AI-driven generative system for 3D-printed footwear. Core components include a high-precision foot scanning app, an AI engine for parameter-driven shoe last generation based on foot morphology, an automated Rhino workflow for upper and sole optimization, and an AI algorithm for the automatic generation of lattice structures from shoe models. Paper submission is planned in the near future.
Alongside my studies at Monash University, I also delivered several AI Agent projects for AU local businesses owners or Github community, such as the voice customer services agent Ringbot, the open source learning platform Learnify, a customized fire safety report drafting agent, among others.
Led a 22-member technical team at the R&D Center, reporting directly to the CTO. Implemented agile methodologies, built complete DevOps workflows, developed core features, and introduced AI-assisted programming, achieving over 100% increase in development efficiency and 85% reduction in customer-reported issues.
I contributed to the development of an independently developed PLM system called Wonder Insight, where I played a key role in the development of many core functions. I also participated in the development of many Siemens Teamcenter projects.
Introduction: An AI-powered TRPG game where you can embark on exciting adventures with your favorite characters. Through a structured engine, the game organically integrates a combat system, an AI GM adventure system, a quest system, an equipment system, and AI-driven dialogue. It also asynchronously generates high-quality story CGs and narrative videos. What players experience is not merely simple character-card conversations, but a vivid world, a cast of well-developed companions, and endless possibilities.
My Contribution: I built the core architecture of the project, the AI orchestration system, all 3D-related features, all DevOps pipelines, and the highly available production Agent sandbox.
Tech Stack: React, Three.js, Gaussian Splatting, Unity, Firebase, MCP, ACP orchestration
Introduction: An AI-powered platform that transforms multi-format content (PDF, Doc, Image, Youtube, etc.) into interactive learning materials including mind maps, notes, flashcards, quizzes, and AI-generated podcasts with real-time 1v1 teaching conversations.
My Contribution: I founded this open-source project and contributed everything, including technical architecture design, feature development, CI/CD pipeline setup, cloud deployment etc.
Tech Stack: Next.js, Prisma, PostgreSQL, GraphQL, LangGraph, Vertex AI, Docker, GCP, Supabase
Introduction: An AI voice assistant for effortless reservations. It features low-latency audio responses, natural-sounding voice, support for complex business logic, and seamless integration with phone calls.
My Contribution: Built the prototype with a small team and sold it to a Melbourne startup. Responsible for low-latency audio-to-audio AI, A2A workflow, RAG implementation, and DevOps.
Tech Stack: React, webRTC, OpenAI Realtime, LangGraph, Cloudflare, Supabase, PostgreSQL, GraphQL
Introduction: A cross-platform 3D model viewer for manufacturing with JT format support. Key features include parameter/annotation preservation, efficient BOM-based multi-file rendering (90% time savings), dynamic hot-reload updates, and virtualized rendering for giant models with 100,000+ parts.
My Contribution: Sole developer responsible for the complete architecture and implementation. Served as primary technical liaison with Siemens support team.
Tech Stack: Three.js, WebGL, Vue, Siemens JT Toolkit, Siemens Vis Web, Docker
Introduction: A self-developed Product Lifecycle Management (PLM) platform. Benchmarking against Siemens Teamcenter, it covered over 70% of the features and in most features it's faster than Teamcenter. It earned over 10 million RMB in enterprise contracts within the first year after launch.
My Contribution: I'm responsible for the technical architecture, system design, and development of the core features. I also in charge of the agile iteration, weekly public release, and delivery management.
Tech Stack: Micro-frontend, SpringCloud, MySQL, Redis, RabbitMQ, Kafka, Docker, Kubernetes, Jenkins