Portrait of Buddhi Ashan

Buddhi Ashan Mallika Kankanamalage

Postdoctoral Fellow, Department of Computer Science · The University of Texas at San Antonio

I redesign classical geometric data structures making them efficient in theory and fast in practice on parallel hardware. On top of them I build fast retrieval systems that search billions of shapes over large spatial data. These systems can extend naturally into metric-native retrieval for AI systems.

I work on parallel algorithms and data structures over large geospatial datasets: polygon clipping harnessing GPUs that handle degenerate cases, which are common in real-world data, improve non contributing line segments employing segment trees, and quadtree-based encodings that make shape-based similarity search practical at the scale of a billion polygons. My work has appeared at CCGrid (best paper finalist), ICPP, IEEE BigData, and ACM SIGSPATIAL, and have released code and benchmark datasets publicly. Additionally, I lead Research Activities Management Portal (RAMP) development for the NSF-funded ScooterLab project.

Education

Appointments

  • Postdoctoral Fellow Department of Computer Science, UT San Antonio Sep 2024 – present
  • Graduate Research / Teaching Assistant Department of Computer Science, UT San Antonio 2019 – 2024
  • Graduate Assistant Department of Computer Science, Sam Houston State University 2017 – 2019
  • Temporary Lecturer / Demonstrator University of Kelaniya, Sri Lanka 2016 – 2017

Selected publications

All publications →