CS student at Beijing Institute of Technology — studying the brain and deep learning together.
I work on continual learning and personalization, with a belief that real intelligence adapts and grows with its context rather than being frozen at training time. Currently building a continual personalization framework for LLMs, growing toward reinforcement learning and biologically plausible adaptation.
- Continual personalization in LLMs — how a model adapts to a specific user over time without forgetting general capabilities
- Personalized safety guardrails — user-specific protective boundaries shaped through continual adaptation
- Human state modeling — physiological and behavioral signals as input to adaptive AI systems
pichu — Terminal-native coding agent with composable tools, sub-agents, and persistent memory.
Website: https://yeabwang.github.io/ LinkedIn: https://linkedin.com/in/yeabsira-tesfaye