LAS VEGAS: Amazon Web Services (AWS) today announced that Mesolitica, a Malaysian startup specializing in training large language models (LLMs), has built a Malay-language generative artificial intelligence (gen-AI) LLM called MaLLaM on its cloud services. The MaLLaM model is designed to understand local nuances, including slang and colloquialisms that combine different dialects, Bahasa Melayu, and 16 other regional languages, for use in AI-assistants across various industries.
According to BERNAMA News Agency, AI-assistants powered by MaLLaM can offer quick, accurate responses to citizens’ inquiries in multiple languages, including dialects from Malaysian states such as Johor, Kedah, Sarawak, Selangor, and Terengganu. This innovation aims to enhance citizen communication and data processing capabilities across Malaysia’s culturally diverse landscape, said AWS.
AWS ASEAN managing director Jeff Johnson highlighted that Malaysian enterprises implementing MaLLaM could improve operations with generative AI in r
egional languages, thereby assisting underserved audiences like farmers in rural areas. This technology enables them to make data-driven decisions using real-time parameters. Furthermore, the Malaysian government is contemplating integrating MaLLaM into its operations, aligning with the nation’s broader objective of AI sovereignty and local data governance, Johnson mentioned during a media briefing for ASEAN countries at AWS re:Invent 2024.
According to AWS, Mesolitica is an active participant in the AWS Asia-Pacific and Japan Generative AI Spotlight Programme. This four-week accelerator programme is designed to support early-stage startups in the region focusing on developing generative AI applications. Mesolitica is also one of two Malaysian companies that have received AWS credits from the AWS Activate Programme, a comprehensive initiative offering resources such as AWS credits, technical support, training, and tools to help startups build, launch, and scale their applications on AWS.
“Mesolitica has joi
ned the AWS Partner Network, a global programme created by AWS to assist businesses in leveraging AWS for growth and success,” according to a statement by AWS. The company has notably improved its machine learning operations by utilizing AWS services, migrating its model training workloads to Amazon Elastic Cloud Compute (Amazon EC2), and deploying inference workloads using Amazon EC2 G5 instances. These instances provide cost-effective GPU (graphics processing unit) acceleration for AI models.
The AWS re:Invent 2024 event, held from December 2-6, 2024, at various venues across Las Vegas, is a learning conference hosted by AWS for the global cloud computing community. The in-person event features keynote announcements, training and certification opportunities, access to over 2,000 technical sessions, the Expo, and after-hours events. It is anticipated to attract approximately 60,000 attendees from around the world.