Localising LLMs With Mesolitica’s MaLLaM

On this episode of Tech Talk, we’re joined by Khalil Nooh, Co-Founder and CEO of Mesolitica, a Malaysian startup at the forefront of developing large language models (LLMs), who have recently built MaLLaM, a generative AI LLM specifically designed to understand and process local linguistic nuances like slangs, colloquialisms, and diverse dialects. The LLM is focused on Bahasa Malaysia and 16 other regional languages which is said to transform industries across the country.We’ll dive into the technical intricacies behind MaLLaM, which has been trained on nearly 200 billion tokens from 197 datasets of Malay-specific content. Khalil will also discuss how this powerful LLM is already revolutionizing AI applications in customer service, content generation, and data analysis. Plus, we’ll explore the challenges of embedding cultural relevance into AI systems, and Mesolitica’s exciting plans for the future as they continue to enhance MaLLaM.See omnystudio.com/listener for privacy information.