Max Robotics — Global Service Robot Directory + US Certification

About

Haziq Razali is a researcher at the forefront of human-robot interaction and motion forecasting, specializing in how machines can understand and replicate complex human actions. His work centers on generating and forecasting bimanual object manipulation sequences—a critical capability for collaborative robots and augmented reality systems. Razali’s major contributions include pioneering methods for action-conditioned generation of two-handed object interactions, a relatively new problem that moves beyond whole-body motion without object contact. He has also developed innovative approaches using eye gaze to forecast human pose during everyday pick-and-place actions, demonstrating that gaze is a powerful indicator of intent for robots operating alongside people. His research extends to modeling human-to-human object handovers through multitask variational autoencoding, enabling assistive robots to replicate this joint action. With papers published in 2021–2024, Razali’s work has already garnered citations (e.g., 6 citations each for his 2022 and 2023 papers), reflecting its growing impact. Notably, his 2024 work on forecasting bimanual sequences from unimanual observations tackles the challenging task of inferring missing arm motion, pushing the boundaries of predictive modeling for assistive technologies.