Seoul, KR · Sookmyung Women's University
AI/ML researcher working on modular RAG pipeline diagnostics — bridging research and product engineering.
// locating which module actually bottlenecks the pipeline, not just the average gain
KDD '26
Split-based analysis that identifies which stage of a modular RAG pipeline — retrieval, reranking, or generation — is the true bottleneck, rather than relying on average oracle gain alone.
CIKM Short Paper (extension)
A diagnostic method for modular RAG pipelines using gain vectors, incidence maps, and a split-conditioned diagnostic operator, extending the KDD submission's bottleneck analysis.