Kay Zhu is a software engineer working on 3D reconstruction from geo imagery at Google. His current main research interest is the intersection of multiview geometry, multiview stereo, machine learning and 3D graphics.
Previously he worked on XLA TPU compiler for TensorFlow at Google Brain and Platforms. Before that he worked on the light field project in Matt Pharr’s group in Google Daydream, and before that he worked on a low-latency, real-time and distributed software-based live video transcoding pipeline and its playback component at Google Fiber.
He occasionally refers himself in third person.