Join Pratyusha Mukherjee at SuperAI Singapore 2025 as she explores the vital theme of integrating cultural context in building Large Language Models (LLMs). Her session addresses the pressing issue of AI's uneven impact, particularly in Southeast Asia, and presents a compelling case for including diverse cultural nuances in AI development to ensure equitable benefits across low-resource regions.

​

Pratyusha emphasizes that while AI holds incredible potential as a multiplier in various sectors, millions are being left behind due to linguistic and cultural gaps. In this insightful presentation, she outlines her collaborative efforts with partners across Southeast Asia to mitigate these challenges. Discover the transformative applications of AI in supporting migrant workers, small-scale farmers, and educational systems in Southeast Asia, highlighting AI's capacity to address inequalities and empower communities.

​

Key highlights include the development of tailored data sets through the innovative use of image prompts and community engagement, the groundbreaking launch of Aquarium—a beta platform enabling regional developers to access contextual data sets, and the introduction of the Sea Helm leaderboard for evaluating AI performance across Southeast Asian languages.

​

Gain insights into cutting-edge model training techniques like the KM method developed at Google DeepMind, designed to optimize LLMs without loss of foundational skills. Pratyusha also discusses the future direction of the project, focusing on multimodality, cultural safety, and value alignment.

​

Watch to learn how openness, adaptability, and strategic partnerships are paving the way for AI to become a universal tool for empowerment. We invite you to like, comment, and subscribe to our channel to stay updated on more thought-provoking sessions from SuperAI Singapore 2025.