About Me

I’m a Ph.D. candidate in Computer Science at the University of South Carolina. I build AI systems that can reason and plan more reliably, especially when the output must follow rules, stay consistent across steps, or use trusted background knowledge.

My research combines three pieces:

• Large language models (LLMs) for flexible reasoning and language understanding
• Structured knowledge (knowledge graphs) to ground decisions in explicit facts and relationships
• Learning-based search (deep reinforcement learning) to generalize problem-solving strategies, especially for pathfinding and planning

Recently, I interned at IBM Research on a fast–slow architecture with a metacognitive feedback loop that improves LLM reasoning performance and efficiency. I also develop foundation-model-style approaches for pathfinding, and knowledge-grounded retrieval tools that make tasks like systematic literature reviews faster and more reproducible.

My work has appeared in venues including AAAI, ICAPS, IEEE Big Data, and IEEE Intelligent Systems. I enjoy building end-to-end prototypes, and I’m always open to collaborations on AI for reasoning, planning, and decision support.