Richard Shan

Bio

Richard Shan is an AI researcher focused on mechanistic interpretability and retrieval-augmented generation. His work centers on building pipelines to evaluate and optimize reasoning in large language models, including novel architectures for feature identification and correctness evaluation. He develops applied systems that combine AI with linguistics, such as a modular refreshable Braille device and computational methods for reconstructing ancient languages. Outside of technical work, Richard enjoys playing frisbee.