I research AI, large language models, and agent systems.
I explore how models reason, use tools, and act across modalities. Through Sage and LingModel, I connect research questions with working systems: agent execution, persistent context, and multimodal architectures.
| Project | Focus |
|---|---|
| Sage | An open-source agent framework for complex tasks. My current focus is v2: the runtime, tools, and the applications built around it. |
| NowHow | A desktop learning agent built with Flutter and Sage. Turns real tasks into guided practice, interactive explanations, and feedback. |
| LingModel | Experiments in multimodal model architecture and training, with separate engineering and research tracks. |
- Agent infrastructure — execution lifecycles, tool interfaces, context, and persistent sessions.
- Large language models — model architecture, training, context, and reasoning.
- Multimodal systems — connecting language, visual generation, and agent behavior.