I am a developer and researcher focusing on full-stack web systems, artificial intelligence, and large-scale simulations. I enjoy designing scalable architectures, experimenting with real-time engines, and applying machine learning to practical domains.
- 💻 Skilled in Vue 3 + Vuetify, Python (FastAPI, Django), MongoDB/ScyllaDB, and real-time systems
- 🔬 Built and maintained smart healthcare platforms for dementia care, serving multiple patients and families
- 📚 Exploring Chinese pronunciation correction with AI and tick-based MMORPG simulation (PRIMIS)
- 🤝 Experienced in agile collaboration, system deployment, and server management
- Implemented a fully connected neural network from scratch (NumPy), achieving >98% MNIST accuracy.
- Build My Waifu (Interactive Desktop App)
- Integrated AI dialogue with Live2D models for synchronized facial expressions and animations. Completed ahead of similar market products; handled architecture, Electron/Live2D integration, and release flow.
- https://github.com/BuildMyWaifu/WaifuOnDesktop
- Smart Dementia Care Platform
- Delivered a production-ready system integrating clinical data, caregiver dashboards, and wearable devices. Enabled real-time monitoring, alerts, and case management used across multiple households.
- PRIMIS (Massive Simulation Experiment)
- Developed a tick-driven world engine with scalable backend (ScyllaDB, WebSocket, FlatBuffers). Focused on high-frequency entity updates, performance optimization, and narrative integration.
- AI/ML Coursework & Research
- Experimented with GANs, LLaMA fine-tuning, and reinforcement learning pipelines for 2D environments.
- Frontend: Vue 3, Vuetify, TypeScript, PixiJS
- Backend: Python (FastAPI, Django, asyncio), Node.js
- Database: MongoDB, ScyllaDB, Redis
- DevOps: Linux, Docker, Nginx, Cloudflare, AWS
- AI/ML: PyTorch, Hugging Face, GANs, LLM fine-tuning
- 🌐 Website: kulimi.tw
- 🔗 GitHub: github.com/Xialai-Kulimi
⚡ Always learning, always building.