YTL AI Labs Teams Up With NVIDIA To Build Dataset For Malaysia-Focused AI

YTL AI Labs has collaborated with NVIDIA to develop Nemotron-Personas-Malaysia, an open dataset containing over 1.35 million synthetic personas. This dataset aims to assist developers in creating AI models that accurately reflect Malaysia's diverse population. The personas are based on official Malaysian demographic and labor statistics and were generated from 150,000 base records from sources like the Department of Statistics Malaysia, OpenDOSM, MyCensus, eStatistik, and national labor-force publications.

The dataset does not include any personal data that could link back to real individuals. Each synthetic persona represents a hypothetical individual and includes demographic characteristics like age, gender, ethnicity, occupation, and region, as well as personality traits based on the OCEAN psychological model. It covers major ethnic groups in Malaysia and various religions, occupations, and regional backgrounds, providing a comprehensive representation of the country's population.

The dataset consists of 39 data fields per persona, enabling developers and researchers to test and train AI systems across various scenarios. Notably, Nemotron-Personas-Malaysia is the first dataset in NVIDIA's Nemotron-Personas collection to use Bahasa Melayu as its primary language. This collection also includes datasets for countries such as the US, Japan, India, Singapore, Brazil, France, South Korea, El Salvador, Vietnam, and Belgium.

YTL AI Labs CEO Foong Chee Mun emphasized that AI development should focus not only on the size of the models but also on their ability to understand local context and society. The goal is to enable Malaysians to use global technology while retaining control over local context and priorities. Nemotron-Personas-Malaysia is part of YTL's sovereign AI efforts and is available publicly through Hugging Face for developers, researchers, and enterprises.