About

Hi, I'm Seungkyu Lee, an incoming Ph.D. student in Computer Science and Engineering at Penn State University (Fall 2026). My research focuses on natural language processing, with an emphasis on reliable, efficient, and personalized AI agents.

My interest in AI agents began while developing an information-seeking assistant for over 10 million users at a Gen-AI startup. When given a user query, the model inferred the underlying intent and autonomously invoked external search tools to produce more specific and reliable answers. This experience reshaped my view of language models: not just as text generators, but as bridges between human intent and machine execution.

At SNU's Human-Oriented Language Intelligence Lab, I worked on making agents reason and act more intelligently. In my TACL paper, I built parameter-level API graphs that capture dependencies between tools, enabling more structured and efficient tool use. In my NeurIPS workshop paper, I designed mechanisms that detect and halt redundant reasoning, helping small reasoning models act decisively without overthinking.

Looking forward, I aim to build agents that are both efficient and adaptive—capable of planning tool calls with minimal overhead while understanding user intent and preferences in ambiguous contexts. By combining structured reasoning with personalized behavior, I hope to advance toward AI systems that perform accurately and cost-effectively while collaborating naturally with humans.