Hello world!
I am the Lead ML Engineer at ParadigmAI and a visiting researcher at KIXLAB, the human–computer interaction lab at KAIST. I graduated from UCSC with an honours degree in Computer Science and am currently reading for an M.Sc in Data Science & AI at the University of Moratuwa.
My research interests primarily lie in the intersection of reinforcement learning and computer games. I am particularly interested in studying about intelligent agents that improve with experience.
In a much broader sense, I like to work on theoretical and application research topics involving artificial intelligence.
(+94) 71 908 4020
oshan [DOT] ivantha [AT] g**** [DOT] com
Research Interests
Reinforcement Learning | Machine Learning | Artificial Intelligence | Game Theory | Computer Games | Computer VisionNews
Jul 2026
My paper Affordable Winner's-Curse Correction: Paired-Offset Reevaluation for Hyperparameter Search on a Single GPU was accepted at the GlobalSouthAI workshop at IJCAI-ECAI 2026, held in Bremen, Germany.
My paper Affordable Winner's-Curse Correction: Paired-Offset Reevaluation for Hyperparameter Search on a Single GPU was accepted at the GlobalSouthAI workshop at IJCAI-ECAI 2026, held in Bremen, Germany.
Sep 2025
Our paper On Privacy-Preserved Machine Learning Using Secure Multi-Party Computing: Techniques and Trends was published in Computers, Materials & Continua (CMC).
Our paper On Privacy-Preserved Machine Learning Using Secure Multi-Party Computing: Techniques and Trends was published in Computers, Materials & Continua (CMC).
Aug 2024
Our paper on the Aya dataset — an open-access collection for multilingual instruction tuning — was published at ACL 2024.
Our paper on the Aya dataset — an open-access collection for multilingual instruction tuning — was published at ACL 2024.
Oct 2023
Our paper Detecting Water In Visual Image Streams Captured From Unmanned Aerial Vehicles was published in the Journal of Visual Communication and Image Representation (JVCI).
Our paper Detecting Water In Visual Image Streams Captured From Unmanned Aerial Vehicles was published in the Journal of Visual Communication and Image Representation (JVCI).