Welcome to SEU-PAISys! We are a research group at Southeast University exploring efficient, evolvable, and deployable AI systems.
🙋♀️ What are we working on?
- Embodied AI memory and self-evolution: helping embodied agents accumulate experience, maintain long-term memory, and improve continually through interaction.
- On-device AI infrastructure: making AI inference, adaptation, and resource management efficient across heterogeneous robots and edge devices.
🚀 Featured project
A portable C++ inference runtime for embodied AI models, including vision-language-action models and world-action models. Embodied.cpp provides a shared deployment path across CPUs, GPUs, NPUs, robots, and simulators.
- 💻 Source code
- 🤗 Models
- 📄 Preprint
🌈 How can you get involved?
We welcome research collaborations and open-source contributions. You can open an issue to discuss an idea, report a problem, or propose support for a new model or hardware backend. Pull requests are also welcome—please check the documentation and roadmap of the relevant repository before getting started.
👩💻 Useful resources
🍿 Fun fact
We care about both sides of embodied intelligence: how agents learn and evolve, and how to make them run efficiently on real devices.