Summary
Dinuka Abeywardena is a Perception Lead and computer vision expert with 13+ years of experience applying both geometric and learning-based methods to real-world robotics and aerial autonomy. At Wing (Alphabet) he led development of production navigation, large-scale 3D reconstruction, DSM pipelines, and a vision-only localization system that now supports hundreds of thousands of commercial missions and thousands of UAVs. He blends PhD-level robotics research with pragmatic engineering, having moved novel algorithms into high-throughput production systems that process millions of images weekly. Dinuka has driven evaluation and adoption of learned features (SuperPoint, SuperGlue, LightGlue and custom models) to augment geometric pipelines, and he’s particularly focused on where learning complements — rather than replaces — classical vision. Based in Mountain View, he brings cross-functional leadership experience from academia to Google and Wing, with a knack for turning safety and business requirements into scalable perception solutions.
13 years of coding experience
4 years of employment as a software developer
Bachelor of the science of Engineering, Electronic and Telecommunication, Bachelor of the science of Engineering, Electronic and Telecommunication at University of Moratuwa
Doctor of Philosophy (Ph.D.), Robotics, Doctor of Philosophy (Ph.D.), Robotics at University of Technology, Sydney