Become a sponsor to Bo Liu
👋 Hello there!
I'm Bo Liu, a Ph.D. student in the Department of Computer Science at the National University of Singapore. My research focuses on the intersection of Reinforcement Learning (RL), Reasoning, and Machine Learning (ML) Systems, with an emphasis on their applications in complex, real-world environments.
🧠 A Few Highlights of My Work
I've been fortunate to contribute to several exciting projects, including:
- [TorchOpt](https://github.com/metaopt/torchopt): An Efficient Library for Differentiable Optimization, featured as a PyTorch Ecosystem Project and accepted in the Journal of Machine Learning Research (JMLR).
- [EnvPool](https://github.com/sail-sg/envpool): A Highly Parallel Reinforcement Learning Environment Execution Engine, showcased at NeurIPS 2022.
- [DeepSeek-V2](https://github.com/deepseek/DeepSeek-V2): A Strong and Efficient Mixture-of-Experts Language Model, which achieves top-tier performance among open-source models, enabling economical training and efficient inference.
- [DeepSeek-VL](https://github.com/deepseek/DeepSeek-VL): A Vision-Language Model for Real-World Applications, designed for diverse multimodal tasks, achieving state-of-the-art performance across various benchmarks.
🌱 Beyond Research
Outside academia, I'm a passionate soccer fan and enjoy playing in my free time. I also love collaborating with others to explore the possibilities of AI and reinforcement learning in diverse fields. If you’re interested in my work or want to chat about AI, research, or soccer, feel free to reach out to me via [email](mailto:[email protected])!
🎯 Why Your Support Matters
Your sponsorship will help fuel my ongoing research, contributions to open-source projects, and collaborations with talented individuals around the world. Together, we can continue to push the boundaries of AI and machine learning.
Thank you for considering supporting my work – every bit of help makes a difference! 🙏
Let’s explore the possibilities of AI together! 🚀
Featured work
-
sail-sg/envpool
C++-based high-performance parallel environment execution engine (vectorized env) for general RL environments.
C++ 1,115 -
metaopt/torchopt
TorchOpt is an efficient library for differentiable optimization built upon PyTorch.
Python 561 -
deepseek-ai/DeepSeek-V2
DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
-
deepseek-ai/DeepSeek-VL
DeepSeek-VL: Towards Real-World Vision-Language Understanding
Python 2,309