Matthew Zhao

Matt is a software engineer and researcher with experience building production NLU systems and conversational AI at Amazon Lab126, where he contributed to Alexa Conversations and FireOS. His work includes designing and deploying bi-LSTM-CRF pipelines for intent classification and entity recognition, developing end-to-end inference stacks for dialog systems, and exploring POMDP-based dialog simulators that led to a paper accepted at NeurIPS 2020’s Human in the Loop Dialogue Systems Workshop. Beyond large-scale voice AI, Matt has shipped products like FoodWatch, an image-recognition-powered food tagging app backed by Azure Mobile Services, and Drop, an Android platform game prototyped in 24 hours, where he implemented core gameplay mechanics and sensor-driven controls.