I'm a software engineer passionate about building the next generation of AI systems. My work is centered around the post-training of large models and developing robust AI Infrastructure.
I am a strong believer in the power of open source to democratize AI and address the public's concerns about its rapid advancement. I believe that transparent, collaborative development is the key to building safe and beneficial AI for everyone.
I am proud to have contributed to a range of impactful open-source projects across both academia and industry.
- HKUNLP/Dream: The large language diffusion models trained by Hong Kong University NLP Lab.
- thunlp/ProactiveAgent: An agent that actively use tools through captured computer operation signals created by Tsinghua University NLP Lab.
- bytedance/tarsier: A multimodal large language model developed by ByteDance Omni Lab that accurately analyzes images and videos.
- vllm-project/vllm: A high-throughput and memory-efficient inference and serving engine for LLMs.
- openai/triton: An open-source language and compiler for writing highly efficient GPU code for deep learning.
- vllm-project/production-stack: Focused on building a robust, production-ready stack for serving LLMs with vLLM.
- LMCache/LMCache: Redis for LLMs. A project designed to optimize LLM serving by caching KV-pairs.
- huggingface/transformers: A library of pretrained text, computer vision, audio, video, and multimodal models for inference and training.