- 📍 Based in: Shanghai, China
- 🏢 Role: AI Infra Team at 📕 Xiaohongshu
- 🤖 Focus: Post-training of large language models (SFT, RL, alignment, optimization)
- 💻 Stack: Python, PyTorch, Distributed Training, Optimization; Golang, C++, C#, TeX, Elisp…
- 🔬 MLLM post-training techniques, especially large-scale reinforcement learning with multimodal data
- ⚡ Training/Inference acceleration & model efficiency
- 🧩 Improving stability and scalability of AI infrastructure
;; This is my work...
(=> (++ (⚙️ 🐛 ⚡) (🧠 📊 🔍))
(=> (++ 📈 📦)
(🚀 🎉)))- ✨ Deep Emacs enthusiast — I write, organize, and live in Emacs + Org Mode
- ☕ Pour-over coffee lover — exploring beans, refining techniques, enjoying the process with
COMANDANTE::C40. - 🎵 Post-rock music
☕ Smooth is 🚀 fast
Perfect is SHIT
See how to do the nyan on NYAN.CAT!




