Seoyeon Kim
I am Seoyeon Kim, an AI research engineer. My strongest technical depth is in multimodal systems: multimodal LLMs, multimodal RAG, and vision–language retrieval—how text, layout, and images are indexed, retrieved, and fed into models so answers stay grounded in the right evidence. In production I work on LLM agents and RAG: when a question needs the right knowledge source, tool, or retrieval step rather than a single pipeline, I design multi-agent flows with orchestration and tool calling, and tie hybrid search to that execution. I also keep tracing and evaluation in the loop so quality and failures are visible, not only the final answer.